Introduction

An epithelial layer with a mucus-rich surface lines the gastrointestinal tract (or gut) of animals. This dynamic environment is a primary interface between host immunity, symbiotic microbes, and dietary antigens Li et al. (2015); Johansson et al. (2011). During colonization of the gut, bacteria encounter physical, chemical, and biological forces. They must compete for nutrients and niche space while managing the stress of digestive enzymes and host immune factors Duncan et al. (2021); Harrington et al. (2021); Hornung et al. (2018); Moran et al. (2019). Gut-colonizing bacteria also encounter bacteriophages (phages), or viruses that infect bacteria Mirzaei and Maurice (2017); e.g., over 10¹² viruses have been estimated in the human gut Shkoporov and Hill (2019).

Phages display both lytic and temperate lifestyles. While the former impacts bacterial community dynamics through lysis, the effect of temperate phages on bacterial communities, especially those associated with animals, remains poorly understood. Conventionally, temperate phages integrate into bacterial genomes as prophages and remain ‘dormant’ until an external trigger activates them to enter the lytic cycle Lwoff (1953); Boling et al. (2020); Howard-Varona et al. (2017); these bacteria are considered ‘lysogenized’ and referred to as lysogens Lwoff (1953); Howard-Varona et al. (2017). Prophages often encode accessory genes that can influence bacterial traits and behaviors Mills et al. (2013). These genes can encode virulence factors, antibiotic resistance genes, and those that provide superinfection exclusion, thereby protecting their bacterial hosts from infections by related phages Bondy-Denomy and Davidson (2014). Based on the site of integration, prophages can also impact the expression of bacterial genes Aziz et al. (2005). Bacterial lysogens exist in every environment Jiang and Paul (1998); Silveira et al. (2021); Leigh et al. (2018) and are particularly prevalent in the microbiomes of diverse animals Kim and Bae (2018); Shkoporov and Hill (2019).

Prophage induction results in bacterial lysis and can be mediated by various stressors, including antibiotics and inflammatory processes Allen et al. (2011); Banks et al. (2003); Diard et al. (2017); Fang et al. (2017); Garcia-Russell et al. (2009); Maiques et al. (2006); Nanda et al. (2015); Wang et al. (2010); Zhang et al. (2000). Because prophages can influence phenotypes of their bacterial hosts such as biofilm formation Nanda et al. (2015), understanding the impact of lysogens in animal microbiomes is becoming a research priority Fortier and Sekulovic (2013); Hu et al. (2021); Lin et al. (1999). In the gut, biofilms that associate with host mucus may benefit the host by enhancing epithelial barriers against pathogenic invasion Swidsinski et al. (2007). Integration of prophages into bacterial genomes may impart functional changes that could be important in surface colonization. For example, in E. coli K-12, integrating the Rac prophage into a tRNA thioltransferase region disrupts biofilm functions Liu et al. (2015); deleting this prophage decreases resistance to antibiotics, acids, and oxidative stress. Some of these traits may be affected by prophage-specific genes Wang et al. (2010). Prophages have also been shown to influence biofilm life cycles in Pseudomonas aeruginosa Gödeke et al. (2011); Rice et al. (2009).

Lysogenized Shewanella species colonize the gut of Ciona robusta, an ascidian collected in Southern California waters Dishaw et al. (2014b); Leigh et al. (2017) and referred to here as Ciona. Tu-nicates like Ciona are a subphylum (Tunicata) of chordates that are well-established invertebrate model systems for studies of animal development Chiba et al. (2004); Davidson (2007); Liu et al. (2006) and are now increasingly leveraged for gut immune and microbiome studies Liberti et al. (2021). Armed with only innate immunity, Ciona maintains stable gut bacterial and viral communities Dishaw et al. (2014b); Leigh et al. (2018) despite continuously filtering microbe-rich seawater. Previous efforts to define gut immunity in Ciona revealed the presence of a secreted immune effector, the variable immunoglobulin (V-Ig) domain-containing chitin-binding protein, or VCBP, that likely plays important roles in shaping the ecology of the gut microbiome by binding bacteria and fungi (as well as chitin-rich mucus) on opposing functional domains Dishaw et al. (2011); Liberti et al. (2019). VCBP-C is one of the best-studied VCBPs expressed in the stomach and intestines of Ciona, shown to bind bacteria in vitro and in the gut lumen Dishaw et al. (2016, 2011). Based on various in vitro and in vivo observations, it was proposed previously that VCBPs likely modulate bacterial settlement and/or biofilm formation Dishaw et al. (2016); Liberti et al. (2019, 2021). The potential influence of soluble immune effectors on host-bacterial-viral interactions is particularly interesting. However, the possibility that prophages may influence interactions between bacteria and secreted immune effectors like VCBPs remains to be explored.

Shewanella fidelis strain 3313 was isolated previously from the gut of Ciona and found to contain two inducible prophages, SfPat and SfMu Leigh et al. (2017). Furthermore, in vitro experiments demonstrated enhanced biofilm formation in S. fidelis 3313 in the presence of extracellular DNA (eDNA) that may originate from lytic phage activity Leigh et al. (2017). Other Shewanella species have previously demonstrated a link between phage-mediated lysis and biofilm formation Gödeke et al. (2011). For example, in S. oneidensis strain MR-1, Mu, and Lambda prophages enhance biofilm formation via eDNA released during prophage-induced lysis, with genomic DNA likely serving as a scaffold for biofilms Gödeke et al. (2011). Similarly, the P2 prophage of S. putrefaciens strain W3-18-1 influences biofilm formation via spontaneous induction at low frequencies, resulting in cell lysis and contributing eDNA that can mediate biofilm formation Liu et al. (2019).

Here, we set out to isolate and characterize the influence of the prophage, SfPat, on its host, S. fidelis 3313. Since the last description Leigh et al. (2017), the genome of S. fidelis 3313 was improved by combining long-read and short-read sequencing, which resulted in improved resolution of the genomic landscape within and around the prophages. A homologous recombination-based deletion strategy was designed to generate a deletion mutant (i.e., knockout) of SfPat. We report that deletion of SfPat results in reduced bacterial motility and increased biofilm formation in vitro. These changes in bacterial traits and behaviors are associated with the expression of genes regulating important signaling molecules and a corresponding impact on host immune gene expression during gut colonization in Ciona juveniles. Gut colonization experiments in laboratory-reared Ciona juveniles comparing wild-type (WT) and SfPat prophage knockout (ΔSfPat) mutant strains demonstrate that SfPat influences gut colonization outcomes, e.g., niche preference, and these effects are influenced by host gene expression The results reported herein reflect complex inter-kingdom interactions among prophages, bacterial hosts, and animal immune systems.

Results

Sequence verification of prophage deletion mutant strains

Colony PCR and single primer extension sequencing were both used to validate the prophage deletion, using primers EDK81/82 for SfPat (Table 2)(Figure 1 a). All recovered amplicon sizes were consistent with the predictions for SfPat deletion (Figure 1 b). The SfPat deletion (ΔSfPat) strain was named JG3862 (Table 1). Genome sequencing of the deletion mutant strain did not reveal the significant introduction of additional mutations or DNA modifications (Figure 1 b). The WT and Δ SfPat strains were then used for in vitro and in vivo experiments to understand the potential role of prophages in shaping S. fidelis 3313 colonization dynamics in the gut of Ciona.

Table 1.

Table 2.

Prophage deletion modulates biofilm formation and motility in S. fidelis 3313 in vitro

Deletion of SfPat from S. fidelis 3313 contributed to an overall increase in biofilm formation as quantified by crystal violet staining by at least 14% compared to WT strains (Figure 2 a). Conventionally, to form a biofilm, bacteria will settle and initiate stationary growth dynamics Watnick and Kolter (2000). We studied bacterial swimming on simple semi-solid media to determine if the prophages influenced motility and chemotaxis in S. fidelis 3313. Bacterial motility was measured by the spread diameter from a primary inoculation point after overnight incubation (Figure 2 b). The WT strain demonstrated a mean diameter of 8.21 mm, while ΔSfPat resulted in a decreased mean diameter of 4.04 mm, demonstrating reduced motility and increased biofilm formation.

Figure 1.

Figure 2.

The influence of prophages on Ciona gut colonization by S. fidelis 3313

Since swimming and biofilm formation behaviors also depend on quorum sensing mechanisms, changes in the expression levels of four genes regulating cyclic-di-GMP (pleD, pilZ, chitinase and pdeB) were measured by qPCR between WT and Δ SfPat. We found no significant change in the expression of these genes in bacteria recovered from 24-hour biofilms (in vitro) (Figure 2 c and Sup fig s1a). However, when these bacteria were introduced to metamorphic stage 4 (MS4) Ciona for 24 hours, the RT-qPCR revealed significant changes in the expression of pdeB (Figure 2 c and Sup fig s1a). The bacterial gene, pdeB, encodes a phosphodiesterase enzyme that degrades c-di-GMP and can serve as a negative regulator of motility, a positive regulator of biofilm formation, and a quorum sensing signal Chao et al. (2013). Colonizing WT S. fidelis 3313 demonstrated higher pdeB expression than the ΔSfPat mutant strain (Figure 2c).

Swimming and biofilm formation often facilitate bacterial colonization of a host. We investigated whether prophages could impact the ability of S. fidelis 3313 to colonize the Ciona gut. Colonization assays were performed on MS4 Ciona juveniles by exposing animals to either WT or ΔSfPat strains, repeating the experiments six times to account for diverse genetic backgrounds (Figure 3 a and b), i.e., using gametes from distinct outbred adults. After exposure to the bacterial strains, retention was estimated by recovering bacteria from animals and quantifying colony-forming units (cfu) at different time points. The ΔSfPat strain revealed a statistically insignificant 1.3-to-1.5-fold change in retention compared to the WT strain after 1 hour of bacterial exposure, a time point that mimics initial colonization (Figure 3 a). However, after 24 hours of exposure, WT was over two-fold retained in the gut than the ΔSfPat strain(p<0.05) (Figure 3 b).

Figure 3.

To visualize the localization of WT and ΔSfPat mutant strains in the gut, MS4 Ciona juveniles were exposed for one hour to BacLight Green-stained WT and BacLight Red-stained ΔSfPat strain variants and vice versa (Figure 3 c-e). The one-hour time point reflects changes in the initial colo-nization of juveniles. These experiments revealed differential localization to the stomach epithelial folds by the WT and ΔSfPat mutant strains. The WT strain typically prefers to occupy the pyloric cecum and the posterior portion of the esophagus and entrance into the stomach (Figure 3d and e, Supp fig s3a and s3b). The ΔSfPat was seen to colonize the walls of the stomach during co-colonization (Figure 3d). The ΔSfPat mutant was also found to colonize the stomach and intestines and less of the esophagus (Supp fig s3 c and s3d). These studies reveal spatial and temporal differences in colonization by differentially lysogenized strains of S. fidelis 3313.

Host immune discrimination and impact on lysogenized bacteria

Host immunity also plays an important role in shaping gut homeostasis. Distinct microbes and their antigens and/or metabolites can elicit host immune responses (Rooks and Garrett (2016)). To determine if the Ciona immune system discriminates among S. fidelis 3313 strains differing only in the presence or absence of the SfPat prophage, we examined the expression patterns of a secreted immune effector, VCBP-C, among juvenile MS4 during intestinal colonization. Under normal healthy conditions, VCBP-C is expressed and secreted by the gut epithelium and can bind (and opsonize) bacteria within the gut lumen (Dishaw et al. (2011)) and influence biofilms in vitro (Dishaw et al. (2016)). After one hour of exposure to the S. fidelis 3313 WT and mutant strains, changes were detected in the expression of VCBP-C. Up-regulation of VCBP-C was noted when juveniles were exposed to ΔSfPat mutant strains of S. fidelis 3313, compared to the WT strain (Figure 4 a) by qPCR(p<0.05). Expression of other innate immune genes were evaluated after 24hrs of exposure to the strains; however, statistically significant responses were likely obscured by host genetic diversity. (Figure 4 b).

Figure 4.

Since the presence of SfPat influences host VCBP-C responses, we investigated whether the binding of VCBP-C to bacterial cell surfaces could influence prophage gene expression. S. fidelis 3313 was grown in MA in vitro for 24 hours in the presence or absence of 50μg/ml of recombinant VCBP-C(Dishaw et al. (2016)). RNA was extracted from the biofilms, and gene expression among SfPat open reading frames was monitored, as well as the SOS response regulators, lexA and recA, in S. fidelis 3313. Interaction of VCBP-C with the WT strain was found to suppress the expression of the structural phage protein P5 of SfPat (Figure 5). It was noted that VCBP-C did not significantly alter the expression of lexA and recA, indicators of the SOS pathway (Figure 5). This reveals that VCBP-C binding to the surface of the S. fidelis 3313 strain may not influence prophage structural genes via conventional SOS responses.

Figure 5.

Discussion

In this report, we utilize a phage-deletion strategy to study the influence of a prophage (SfPat) on a gut symbiont (S. fidelis 3313) of the model invertebrate, Ciona robusta. Deletion of SfPat prophage from S. fidelis 3313 revealed phage-mediated disruption of colonization dynamics. We find that the SfPat prophage increases swimming behavior while reducing biofilm formation in the WT. These different bacterial phenotypes influence host immune responses in a manner consistent with influences on gut colonization dynamics. Our identification of a prophage that interferes with biofilm formation in a Shewanella strain contrasts other reports, which show that presence of prophages helps in biofilm formation by lysis and subsequent eDNA in the biofilms. (Gödeke et al. (2011); Liu et al. (2019)). Our study is the first to implicate phage-included change in motility for Shewanella spp, and it indicates that prophages can impact their hosts by modulating traits through diverse mechanisms. For example, some prophages have also been shown to also influence motility in Pseudomonas aeruginosa (Tsao et al. (2018)).

SfPat in S. fidelis 3313 appears to have a variable influence on bacterial retention in the gut of Ciona. As feeding is initiated in MS4 Ciona juveniles, food (or bacteria in the environment, natural or artificially introduced) accumulates in the gastrointestinal tract and on average takes about 45 minutes to begin exiting the anus and atrial siphon as fecal pellets. Thus, we monitored and compared transit and retention of introduced bacteria in the gut of MS4 juveniles at one and 24 hrs after introduction. SfPat deletion reveals a prophage-mediated influence on gut localization and retention. For example, within one hour of exposure, fecal pellets begin to form in the stomach that are enriched for the WT S. fidelis 3313; however, the WT strain then appears to be retained in the posterior portion of the esophagus (just before the entrance into the stomach) as well as in the pyloric cecum (which is a small outpouching just ventral and posterior to the stomach). However, the ΔSfPat mutant strain appears to adhere and colonize the stomach folds and the intestines and not the esophagus. The overall retention of the two strains did not vary significantly within one hour of colonization. However, after 24hrs, the retention of the WT strain was significantly greater than that of ΔSfPat despite the continued feeding behavior of the MS4 juveniles. The increase in pdeB activity of the WT in vivo (Figure 5) might indicate that the secondary messenger, cyclic di GMP, may be degraded in the WT, leading to increased biofilm formation and reduced motility(Jenal and Malone (2006)). Thus, the presence of SfPat influences niche preference and retention.

In addition to host genetics obscuring the influence of phages on colonization of the gut, other biophysical factors that include host immune effectors play crucial and often silent roles in influencing bacterial settlement dynamics. For example, human secretory immunoglobulin A (SIgA) has been shown to enhance and often favor settlement of bacteria both in vitro and in vivo (Bollinger et al. (2006); Thomas and Parker (2010); Pratt and Kolter (1998); Donaldson et al. (2018)), raising a basic question as to whether this phenomenon is more widespread among other secretory immune effectors present in mucosal environments of animals (Dishaw et al. (2014a)). We speculate that while prophages likely impact the behavior of lysogenized bacteria in ways that can influence colonization dynamics, interaction with VCBP-C on the mucosal surface of the Ciona gut likely further influences settlement behaviors (Dishaw et al. (2016, 2014a)). Importantly, we show here that the influence of the SfPat prophage on bacterial physiology leads to a reduced expression of Ciona VCBP-C in the first hours of colonization (an indicator of initial contact of bacteria to a juvenile gut). It remains to be shown if prophages stimulate the production of a bacterial metabolite with immunomodulatory properties or if the host immune system responds to differences in bacterial behaviors or traits, as suggested in the ΔSfPat deletion mutant.

Metagenomic sequencing of gut microbes from healthy humans has revealed that temperate lifestyles are prevalent among phages from these ecosystems (Minot et al. (2011, 2013); Reyes et al. (2010)), an observation also made in the Ciona gut (Leigh et al. (2018)). Various environmental triggers, such as UV light and mutagenic agents like mitomycin C, have been shown to induce a switch from the temperate to lytic cycle via the SOS response, a cell-wide response to DNA damage that can promote survival (Weinbauer and Suttle (1999)). Since, VCBP-C is an immune molecule in the gut that can interact with bacteria, it could influence prophage induction. However, we find that VCBP-C binding on the surface of WT S. fidelis 3313 leads to a reduction in the expression of an important SfPat structural protein P5, suggesting a limitation in SfPat induction in the presence of VCBP-C. No significant changes in lexA/recA expression were observed upon VCBP-C exposure/binding, suggesting a lack of SOS response when exposed to this immune effector. The various mechanisms by which prophages shape colonization behaviors among gut bacteria of animals remain unclear. While the data reported here are only based on one bacterial strain that colonizes the Ciona gut, we find that WT S. fidelis colonize the gut with reduced activation of VCBP-C gene expression compared to ΔSfPat, a trait that may be important in shaping colonization outcomes. We speculate that these observations are more widely applicable since lysogens are so abundant in animal microbiomes. Under normal conditions, VCBP-C protein is present in copious amounts and tethered to chitin-rich mucus lining the gut, as revealed by immunohisto-chemical staining (Dishaw et al. (2016)). Therefore, overexpression of VCBP-C is not necessarily helpful, and can correspond to the induction of additional inflammatory responses, including an overproduction of mucus. Thus, regulation of the production of additional VCBP-C likely serves important roles influencing colonization dynamics.

Since colonization of animal mucosal surfaces is an ancient process (Dishaw et al. (2014b)), prophages and their integration into bacterial genomes have likely evolved to provide fitness benefits in often challenging environments like the gut lumen. Determining the role of animal immunity and prophages in these exchanges is of broad interest. Immune effectors like VCBPs, which undoubtedly possess broad specificities, can bind a range of bacterial hosts; however, it remains to be shown if they bind lysogenized bacteria with different affinities than their prophage-free counterparts. Prophages can also be induced to generate lytic particles that can influence gut microbiome structure, serving as an indirect form of protection for the host Wang et al. (2010); Barr et al. (2013). Prophages can also contribute to the transfer of virulence factors (Nanda et al. (2015); Wagner and Waldor (2002)). Retention of lysogens may be preferred if the prophages provide competitive fitness and retention in the gut. Since lysogens are integral in animal development, immunity, and metabolism(Fraune and Bosch (2010)), there is a transkingdom interplay required for survival, a snapshot of which is shown here.

Methods and Materials

Culture and growth conditions

S. fidelis 3313 used in this study was originally isolated from the gut of Ciona robusta obtained from Mission Bay, CA, USA, as previously described Leigh et al. (2017). The bacterium was cultured using Difco marine agar 2216 (MA) (Fisher Scientific, Hampton, NH) and marine broth (MB) at room temperature. Subsequent genetic manipulations were performed on strains grown in LB/MB, which consists of a mixture of 75% LB (Lysogeny Broth, Fisher Scientific, Hampton, NH) and 25% MB. Strains are listed in Table 1.

Prophage deletion

SfPat was targeted for deletion from S. fidelis 3313 using homologous recombination methods adapted from Saltikov and Newman (2003) to produce knockout mutant strains. First, a pSMV3 suicide vector Saltikov and Newman (2003) was designed with 700 bp regions corresponding to the upstream and downstream sequence of the prophage (Table 2). These flanking regions were amplified and ligated using overlap extension PCR, then directionally inserted into the vector with the restriction enzymes BamHI and SacI (Table 3) Bryksin (2010). Plasmid conjugation was then performed by inoculating a colony of S. fidelis 3313 into a culture of E. coli containing the desired suicide vector on an LB/MB agar plate for two hours. Illumina sequencing (MiGS, University of Pittsburgh) confirmed the deletion of SfPat and the evaluation of any additional genetic changes or mutations (Figure s1a).

Table 3.

S. fidelis crystal violet biofilm assay

WT and ΔSfPat strains were cultured in MB overnight at RT and then diluted to 10⁷ cfu/ml in MB. The cultures were brought to a 2 ml final volume of MB in 12-well dishes and incubated at RT for 24 hours to examine biofilm development. Each variable was tested in technical duplicate. Biofilms were quantified by crystal violet staining as previously described Liberti et al. (2022). Briefly, supernatants were aspirated after 24-hour incubation, and the biofilms were dried and stained with 0.1% crystal violet for 10 mins. The stained biofilms were then gently washed with deionized water, and the amount of biofilm produced was quantified as the intensity of the stain (oD₅₇₀) after biofilm bound crystal violet was extracted from the biofilm with 30% acetic acid. All biofilm assays were performed at least in triplicates.

Motility assay

Soft-agar overlay motility assays were carried out in 12-well dishes to compare swimming behaviors Wolfe and Berg (1989); Kearns (2010) overnight at RT and then inoculated onto the center of soft agar (containing LB/MB and 0.5% low-melt agarose) and incubated at RT overnight. The results were recorded as the distance traveled (in millimeters) from the inoculation zone. Each variable was tested in duplicate. Two perpendicular distances from the inoculation zone were recorded and averaged for each well.

Ciona mariculture

The in vivo colonization experiments were performed on animals reared under conditions termed “semi-germ-free” (SGF), which include minimal exposure to marine microbes. SGF conditions include animals harvested under conventional approaches Cirino et al. (2002) but permanently maintained in 0.22 μm-filtered, conditioned artificial seawater (cASW), handled with gloves, and lids only carefully removed for water/media changes. cASW is prepared by conditioning ASW made with Instant Ocean™ in an in-house sump-aquarium system containing live rock, growth lights, and sediment from San Diego, California; salinity is maintained at 32-34 parts per thousand (or grams per liter). Compared to germ-free Leigh et al. (2016) or SGF, conventionally-reared (CR) includes a step-up exposure to 0.7μm-filtered cASW that increases exposure to marine bacteria during development. The SGF approach is considered an intermediate method of rearing that includes minimal exposures to microbial signals during development (unpublished observations). The animals were reared at 20 ^°C from larval to juvenile stages. The Ciona were collected from Mission Bay, California in order to produce juvenile organisms for each biological replicate. These wild-harvested animals provide a wider genetic diversity compared to traditional model systems, where genetic diversity has been reduced or eliminated through controlled breeding practices.

Gut colonization assays

Both bacterial strains were grown overnight at RT in MB and diluted to 10⁷ cfu/ml in cASW after repeated washes. Metamorphic stage 4 (MS4) animals reared in six-well dishes in cASW were exposed to 5 ml of 10⁷ cfu/ml bacteria in each well for one hour or 24 hours, respectively. MS4 animals are considered part of the 1st ascidian stages (post-settlement stages 1-6, whereas stages 7-8 and onwards are 2nd ascidian stages and reflect young adult animals). MS4 juveniles can be identified as having a pair of protostigmata, or gill slits, on each side of the animal (Chiba et al. (2004)). These juveniles first initiate feeding via newly developed and opened siphons; before this, the gut remains closed, and the interior lumen is unexposed to the outside world. Following this initial exposure or colonization, for various time intervals, the plates were rinsed multiple times with cASW and replaced with fresh cASW. Ten juveniles were chosen randomly for each treatment, pooled, and homogenized with a plastic pestle; live bacteria were counted by performing serial dilutions and enumerating colony-forming units (cfu) via spot-plating assays (Gaudy Jr et al. (1963); Miles et al. (1938)). Each graphed data represents a biological replicate dataset from genetically distinct/ diverse backgrounds of Ciona (represented by separate live animal collection and spawning events). Statistical significance was calculated using the Wilcoxon t-test by pooling data across six genetically diverse biological replicates. Live bacteria in the gut were visualized using BacLight stains and previously described fluorescently labeled bacteria Moran et al. (2019). For BacLight staining, 1 ml of bacterial cultures were grown overnight at RT, pelleted, washed twice with cASW, and stained with 4 μl of BacLight Red (Invitrogen Cat no B35001) or BacLight Green (Invitrogen Cat no B35000) for 15 mins in the dark. The cultures were stained with alternate dyes in different replicates to get unbiased data from changes in fluorescence. The stained cultures were washed twice with cASW, and then diluted to 10⁷ cfu/ml with cASW. MS4 animals were grown in 6-well dishes were then exposed to 5 ml of this culture. Bacteria in the gut of animals were visualized after one hour on a Leica DMI 6000B stereoscope with a CY5 fluorescent filter for BacLight Red and GFP filter for BacLight Green; and imaged-captured with a Hamamatsu ORCAII camera (model C10600-10B-H) and processed with the MetaMorph 7.10.4 imaging suite (Molecular Devices, Downingtown, PA).

Gene expression studies

To determine if the Ciona innate immune system can recognize and respond to unique mutant strains, which differ only in the presence or absence of prophages, candidate immune response markers were tested using reverse transcription quantitative PCR (RT-qPCR). RNA was extracted using the RNA XS kit (Macherey-Nagel, Cat no 740902) from MS4 Ciona juveniles that underwent gut colonization. Complementary DNA (cDNA) synthesis was performed with oligo-dT primers and random hexamers using the First Strand cDNA Synthesis Kit (Promega Cat no A5000) following the manufacturer’s instructions. The amplification was set with the qPCR kit (Promega Cat no A6000) and carried out on an ABI7500 with an initial melting temperature of 95^°C for 2 mins and 40 cycles of 95^°C for 15 sec and 60^°C for 1 min. The innate immune responses tested and their primers are reported in Table 4. Results from four distinct biological replicates are presented. Each replicate includes pooled Ciona juveniles from at least two wells of a 6-well dish. Ciona actin was referenced as an endogenous control. Data was analyzed using ΔΔCT method Pfaffl(2001) and the ABI7500 software suite. To understand bacterial genes that are differentially regulated due to the presence of prophages, bacterial gene expression was studied in vitro and in vivo. The bacterial strains were cultivated in six-well dishes using the same methodology described for biofilm assays. To understand if the host immune molecule VCBP-C induced prophages, WT was cultured in six-well dishes as described in the biofilm section in the presence or absence of 50 μg/ml VCBP-C Dishaw et al. (2016). After 24 hours, the supernatant was discarded for both experiments, and RNA from the biofilm was extracted using the Zymo Research Direct-zol kit. cDNA synthesis was carried out with random hexamers primers, and qPCR was carried out as described above. Targeted bacterial genes are described in (table 5). Rho was identified as the most stable reference gene using RefFinder, which utilizes Bestkeeper, GeNorm, Normfinder, and comparative ΔΔ CT methods (Figure 5 supplement 1) Andersen et al. (2004); Pfaffl(2001); Vandesompele et al. (2002); Watnick and Kolter (2000).

Table 4.

Table 5.

Statistical analysis and data visualization

Statistical analysis and data visualization were carried out in R, version 4.2 Team (2021). Data were plotted with ggplot 3.3.5 Kassambara (2020); the Beeswarm plot was constructed using ggbeeswarm 0.6 Eklund (2021). Beeswarm plots and statistics for motility assays were calculated using replicate averages Lord et al. (2020). Statistical significance was calculated using ggsignif package 0.6.3 or ggpubr0.4.0 Kassambara (2020); Constantin (2021). If the data were found to be normally distributed by Shapiro’s test, then significance was calculated using an unpaired t-test. The Wilcoxon signed-rank test was used to calculate the significance of non-parametric data.

Author contributions

O.N. designed, executed, and analyzed experiments and wrote and edited the manuscript, S.L.G., N.P., C.G.F.A., F.N, A.L., and E.D.K performed experiments and provided feedback and approved the manuscript, M.N.Y., S.J.L., and B.A.L. performed genome sequence analysis, assembly, and provided feedback and approved the manuscript, M.B. and J.A.G. helped interpret data, provided feedback and approved the manuscript, and L.J.D, helped design experiments, interpret data, and helped write, edit and approve the manuscript.

Funding

This project was supported by NSF IOS-1456301 (L.J.D. and M.B.), NSF MCB-1817308 (L.J.D), NSF IOS-2226050/51 (L.J.D. and J.G.), internal awards from the MCOM Deans Office, Ann, and Andrew Hines Endowed Chair in Molecular Genetics, and the USF Institute for Microbiomes to L.J.D. W also got funding through an NSF GRFP award to B.A.L., and a New Investigator Award to O.N. by USF.

Acknowledgements

The authors acknowledge the expertise of Gary W. Litman and John P. Cannon for expert feedback and guidance on earlier efforts of the project.