Mimicking prophage induction in the body: induction in the lab with pH gradients

Introduction

Exploration of the human microbiome has uncovered a vast diversity of viral species, particularly bacteriophages (see review García-López, Pérez-Brocal & Moya, 2019). Phages are vital members of microbiota; they can transform bacterial communities by predation (Fuhrman, 1999; Suttle, 2005; Clokie et al., 2011) and drive bacterial diversity (Buckling & Rainey, 2002; Rodriguez-Valera et al., 2009; Tellier, Moreno-Gámez & Stephan, 2014; Koskella & Brockhurst, 2014). Phages also can protect the epithelia from bacterial infection (Barr et al., 2013). Core phage communities have been identified within several niches of the human body, and the disruption of these communities has been associated with certain gastrointestinal (Wagner et al., 2013; Manrique et al., 2016) and oral (Ly et al., 2014) symptoms/disease. To date, most human virome studies have focused on phages in their lytic form. However, the community of phages replicating through the lysogenic life cycle is likely far greater. The majority of bacteria within the human body are lysogens harboring phage sequences (prophages) within their genomes, including the gut (Breitbart et al., 2003; Reyes et al., 2010; Minot et al., 2011), the oral cavity (Pride et al., 2012), the skin (Hannigan et al., 2015), and the urinary tract (Miller-Ensminger et al., 2018). Similar observations have been made in the microbiota of other organisms, e.g., the murine gut (Kim & Bae, 2018) and the bee hindgut (Bonilla-Rosso et al., 2020). Recent evidence suggests that the lysogenic life cycle may be the dominant life cycle for phages in many environments, not just niches of the body (Knowles et al., 2016).

Induction is the process of a prophage leaving the lysogenic cycle and entering into the lytic cycle. A variety of environmental stressors can trigger this process, including: the SOS pathway, reactive oxygen species, etc., although this process can also occur “spontaneously” (see review Nanda, Thormann & Frunzke, 2015). In the laboratory, the most commonly used method for induction is Mitomycin C (see protocol Raya & H’bert, 2009). Other common laboratory techniques include UV (e.g., Barnhart, Cox & Jett, 1976; Zhang et al., 2019), antibiotics (e.g., Shaikh & Tarr, 2003; Goerke, Köller & Wolz, 2006; Loś et al., 2009), and H₂O₂ (e.g., Figueroa-Bossi & Bossi, 1999; Martín et al., 2009; Loś et al., 2009; Loś et al., 2010). A few studies have also employed changes in temperature (e.g., Kirby, Jacob & Goldthwait, 1967; Meijer et al., 1998; Lunde et al., 2005; Chu et al., 2011), nutrient availability (e.g., Wilson, Turner & Mann, 1998; Lunde et al., 2005; Williamson & Paul, 2006; Zeaki, Rådström & Schelin, 2015), and pH (e.g., Meijer et al., 1998; Lunde et al., 2005; Wallin-Carlquist et al., 2010; Choi, Kotay & Goel, 2010). With respect to the human body, efforts have begun focusing on chemicals and conditions relevant to the niche from which the lysogens were isolated (e.g., Pavlova & Tao, 2000; Oh et al., 2019). In a recent study, food, chemical additives, and plant extracts were shown to induce phages from bacterial taxa common in the human gut (Boling et al., 2020). Similarly, diet has been linked to induction in the murine gut (Kim & Bae, 2018). While we can speculate that induction occurs within the human microbiome, the rate of induction is unknown. Evidence using a mouse model suggests it is likely to be a frequent event in vivo (Diard et al., 2017).

pH-mediated induction was of particular interest to us as we study the urinary microbiota and our prior work found that >80% of bacteria in the bladder are lysogens (Miller-Ensminger et al., 2018). Lactobacillus species, which lower pH by the production of lactic acid (see review Stapleton, 2016), are dominant members of the bladder’s microbiome (see reviews Brubaker & Wolfe, 2017; Mueller, Wolfe & Brubaker, 2017). The pH of urine in healthy individuals can vary widely (Simerville, Maxted & Pahira, 2005; Neugent et al., 2020), and individuals with Type 2 Diabetes and uric acid stones typically have lower pH values (Pak et al., 2003; Cameron et al., 2006). Thus, we were interested to see if pH values deviating from the “norm” stress urinary bacteria, thus mediating induction. Upon our review of the literature, we found just a few studies that had explored this environmental factor and vary in the methods employed. Here we present a protocol for pH induction. pH is an important environmental stressor to consider beyond just the urinary microbiota. For instance, pH changes due to abiotic and biotic factors have been reported in soil (Obour et al., 2017) and marine and freshwaters (Griffith & Gobler, 2020). As a proof-of-concept, we investigated pH-mediated induction in E. coli isolates from the bladder.

Materials and Methods

Results

The pH protocol detailed in the Methods and represented in Fig. 1 cultures lysogens at different pH values as a means of triggering prophage induction. Lysate is collected from each pH adjusted culture from Step 2, and then filtered and spotted on naïve bacterial lawns in Step 3. In Step 4 of the protocol, lawn clearings are harvested, filtered, and replated with naïve bacteria to identify individual plaques, which are then harvested and filtered again. The protocol includes two optional steps: Step 1, which is not represented in the schematic, and Step 5. These two steps are included for identification of induced phages. Step 1 includes sequencing of lysogen’s genome and prophage sequence prediction. Based upon these predicted sequences, PCR primers can be designed for use in Step 5. While the pH conditions shown in Fig. 1 were selected explicitly for our proof-of-concept assays using E. coli isolates of the bladder, the methodology described can be fine-tuned for lysogens isolated from other niches. pH varies across sites in the human body (Evans et al., 1988; Lambers et al., 2006), providing unique growing environments for microbial communities. Additionally, pH changes have been reported in ecological contexts for soil and water environments. Growing isolates at various pH levels that may be experienced in vivo increases our understanding of phage induction.

To test our protocol, we focused on E. coli isolates from the urinary microbiome. We tested pH values 2, 4, 7, 9, and 11, thus capturing the range of values recorded for urine samples. The experimental protocol detailed in the ‘Methods’ section and illustrated in Fig. 1 was implemented for the 7 bladder E. coli strains listed in Table 1, each of which is predicted to contain several putative prophage sequences.

Three laboratory E. coli strains were used as potential susceptible hosts for induced prophages—E. coli B, E. coli C, and E. coli K-12. Each of the 7 bladder lysogens cultured under each of the 5 pH conditions were spotted onto these three laboratory E. coli strains. Of the 105 spots, 49 produced clearings on a bacterial lawn (Table 2) suggesting that pH induced a prophage or spontaneous induction occurred. No clearings were observed on the lawns for the bacterial cultures grown at pH 11, where reduced fecundity of all 7 E. coli lysogens was observed (Fig. S1).

Each clearing was isolated, as shown in Fig. 1, and pour plates were prepared. Figure 2A shows one of these pour plates. Here the plate illustrates the filtered supernatant of UMB1160 grown at pH 4 and plated on the E. coli K–12 host. Supernatants were confirmed to contain phages via TEM. For example, we observed tailed phages from the lysate of UMB1160 grown at pH 4 (Fig. 2B).

To identify induced phages, we performed PCRs in duplicate on individual plaques harvested from the pour plates. Table 3 lists the phages identified. While only 17 of the 49 plaques were identified by PCR, the results highlight two important observations. First, some phages are only induced as a result of a specific pH. For instance, when UMB0149 is grown at pH 7, only one phage, N47 Cat2, is detected. In contrast, when this strain was grown at pH 4, two other phages were detected: N10 Cat5, which lysed E. coli C, and N97 Cat2, which lysed E. coli B. Second, some phages are induced at multiple pH values. When UMB0923 was cultured at pH 4 and pH 7, N3 Cat4 was detected. Similarly, N5 Cat5 was detected in UMB0934 cultured at both pH 4 and pH 7.

Discussion

The pH protocol presented here provides a method for inducing prophages in the lab. Based upon our proof-of-concept assays using E. coli lysogens, it is quite effective. It is important to note that in patients with a urinary tract infection (UTI) caused be E. coli, urine specimens often had a pH around or slightly below 6 (Lai et al., 2019). Over half of the pH cultures tested here produced clearings on naïve bacteria lawns. Clearings of bacterial lawns were observed most frequently for E. coli isolates grown at pH 4 and 7. Of the 49 observed clearings, we identified 17 phages, matching them to predicted phage sequences via PCR (Table 3). The protocol presented in the Methods can easily be adapted for inducing prophages from other bacterial taxa and isolates from other environments; modifications can include the pH values tested in Step 2 and the culture conditions (temperature, oxygen levels, duration, etc.). These modifications should be made keeping in mind the conditions from which the lysogen was isolated. Thus, selecting pH values that are likely to cause stress provide insight into prophage induction in vivo.

For our interests, culturing bacterial isolates of the bladder at pH values representative of the broad range typically observed, can mimic one aspect of the bladder environment. While we acknowledge that our results (Table 2) are biased for those induced phages capable of infecting and lysing one of the three laboratory host strains used, propagation through E. coli C, B, and K-12 increased the titers of the induced phage and provided visual confirmation of their presence. Inclusion of additional putative hosts may reveal additional phages. Alternatively, the lysate of the pH-altered culture could be directly sequenced thus avoiding the need for a susceptible host bacterium. Furthermore, by performing pour plates, we were able to confirm that the bacterial lawn clearing was the result of phage lysis rather than an antimicrobial produced by the bladder isolate, e.g., a bacteriocin.

Several factors may have contributed to negative PCR results. First, our protocol does not include a complete DNA extraction step, omitted to reduce both cost and time. Rather phage DNA is released by heating the lysate (Step 5 i). It is possible that the phage protein coats were not denatured during this heat treatment. While amending the protocol to include phage lysate DNA extraction may circumvent this issue, the efficacy of DNA extraction protocols varies based on sample type (Kleiner, Hooper & Duerkop, 2015; Castro-Mejía et al., 2015, p.; Verbanic et al., 2019; Burgener et al., 2020; Göller et al., 2020, p.). Second, the induced phage may have a low burst size, below the detection threshold of standard PCR, recently shown to be 32 copies of the genome per 1 µL input DNA (Purcell et al., 2016). Third, the PCR primers may not be targeting the phage genome. Although the PCR primers amplified the bacterial DNA, the bioinformatic prediction of the prophage sequence may be incorrect. Phage sequence predictions tools are constantly improving. It remains a challenge to discern between those prophages capable of entering the lytic cycle and those that cannot. Whole genome sequencing of the lysate produced at the end of Step 4 of the protocol would provide insight into the accuracy of these prophage sequence predictions.

We did not observe clearings on E. coli isolates grown at pH 11, and only two isolates grown at pH 2, UMB0527 and UMB1135, produced clearings on the spot plates. At these extreme pH values, we noticed reduced growth of the bacterial cultures (Fig. S1), which may have prohibited prophage induction or phage propagation. While some phages were only identified from a culture for a single pH condition, others were identified in the lysate of cultures from multiple pH levels. This may reflect a range of pH values able to induce the phage or it may simply be the result of spontaneous induction. Spontaneous induction has been observed in isolates from the urinary tract, such as Streptococcus anginosus (Brassil et al., 2020).

Conclusions

The pH protocol is efficient in the induction of prophages from E. coli isolates from the bladder. The range of pH values observed in the urine may be the result of differences in microbial community structures and phage induction may be a contributing factor. Applying stressors that are directly applicable to the urinary tract provides insight into how phages may shape the bladder microbiota. Similar strategies can be employed for other anatomical niches. While here we have focused our efforts on induction by pH, other conditions of the bladder could also be explored. Artificial urine media (e.g., Brooks & Keevil, 1997; Christmas et al., 2002; Khan et al., 2017) are limited in the number of bladder bacterial taxa they can support. As such there are several growth factors that have yet to be determined that may also be triggering or suppressing induction within the urinary microbiota. The protocol presented here considers one well-studied aspect of the bladder environment and provides a means of replicating what might be a natural process in the urinary tract.

Supplemental Information