Small Alarmone Synthetases RelP and RelQ of Staphylococcus aureus Are Involved in Biofilm Formation and Maintenance Under Cell Wall Stress Conditions

The stringent response is characterized by the synthesis of the alarmone (p)ppGpp. The phenotypic consequences resulting from (p)ppGpp accumulation vary among species, and for several pathogenic bacteria, it has been shown that the activation of the stringent response strongly affects biofilm formation and maintenance. In Staphylococcus aureus, (p)ppGpp can be synthesized by the RelA/SpoT homolog Rel upon amino acid deprivation or by the two small alarmone synthetases RelP and RelQ under cell wall stress. We found that relP and relQ increase biofilm formation under cell wall stress conditions induced by a subinhibitory vancomycin concentration. However, the effect of (p)ppGpp on biofilm formation is independent of the regulators CodY and Agr. Biofilms formed by the strain HG001 or its (p)ppGpp-defective mutants are mainly composed of extracellular DNA and proteins. Furthermore, the induction of the RelPQ-mediated stringent response contributes to biofilm-related antibiotic tolerance. The proposed (p)ppGpp-inhibiting peptide DJK-5 shows bactericidal and biofilm-inhibitory activity. However, a non-(p)ppGpp-producing strain is even more vulnerable to DJK-5. This strongly argues against the assumption that DJK-5 acts via (p)ppGpp inhibition. In summary, RelP and RelQ play a major role in biofilm formation and maintenance under cell wall stress conditions.

There is some evidence that the stringent response might trigger biofilm formation in S. aureus based on the observation that treatment with mupirocin (Sritharadol et al., 2018;Jin et al., 2020) or serine hydroxamate (De La Fuente-Nunez et al., 2014) results in increased biofilm formation. Moreover, the anti-biofilm peptides IDR-1018 and DJK-5 have been suggested to directly interact with (p)ppGpp, preventing its signaling effects and, thus, biofilm formation (De La Fuente-Nunez et al., 2014. Here, we aimed to investigate the role of the stringent response mediated by the two small alarmone synthetases RelP and RelQ in biofilm formation. Under cell wall stress conditions induced by a subinhibitory vancomycin concentration, both RelP and RelQ are crucial for biofilm development. Moreover, (p)ppGpp synthesis prevents biofilm eradication by vancomycin. (p)ppGpp-mediated biofilm formation was shown to be independent of the major stringent response mediator CodY and the main biofilm regulator Agr. The anti-biofilm peptide DJK-5 could prevent biofilm formation of wild type and (p)ppGpp-defective mutants. However, the (p)ppGpp 0 strain is even more vulnerable to DJK-5. Thus, DKJ-5 does not act on biofilm formation via (p)ppGpp inhibition.

Bacterial Strains and Growth Conditions
The strains and plasmids used in this study are listed in Supplementary Table S1. For overnight cultures, the strains were grown with appropriate antibiotics (10 µg/ml erythromycin, 100 µg/ml spectinomycin) at 37 • C and 200 rpm. For biofilm analyses, the following media were used: chemically defined medium (CDM; Pohl et al., 2009)

Strain Construction
For relP complementation, the pCG833 plasmid was constructed. The complete relP operon with its native promoter was amplified by PCR using the primers pCG833gibfor and pCG833gibrev and cloned via Gibson assembly into the BamHI-digested integration vector pCG3. The oligonucleotides used in these procedures are listed in Supplementary Table S2. Due to showing toxicity in E. coli, the plasmid was directly transformed into S. aureus Cyl316 by electroporation. The integration of the plasmid into the genome was verified by PCR using the scv1, scv21, pCG3intfor and pCG3intrev oligonucleotides followed by sequencing. The integrated plasmid was transduced into the target strains using 11 phage. For relQ complementation, the pCG216 plasmid was constructed. relQ with its native promoter was amplified by PCR with the primers BamHrelQkompl-for and BamHrelQkomp-rev and ligated with the BamHI-digested pCG3 vector backbone. The plasmid was verified by PCR with relQdigfor and relQdigrev. The plasmid was transformed into Cyl316 and verified by PCR using the primers relQfor and relQrev for integration of the plasmid. The integrated plasmid was then transduced via phage 11 in the target strains.

Colony-Forming Unit (CFU) and Minimal Inhibitory Concentration (MIC) Determination
For CFU measurements, biofilm-grown bacteria (24 h, 37 • C) were resuspended by thorough pipetting. The bacterial suspension (biofilm resolved and planktonic) was serially diluted in phosphate-buffered saline (PBS), and 10 µl aliquots were spotted onto TSA plates for CFU determination. The MIC was determined by serial microdilution and E-tests.

Biofilm Assay
For the static biofilm assay, 1 ml medium was inoculated in a 24-well polystyrene cell culture plate (Greiner) to obtain an OD 600 of 0.05. After 24 h of static incubation at 37 • C, the wells were washed twice with 1 ml phosphate-buffered saline (PBS, pH 7.4, Gibco). The biofilm was dried at 40 • C for 30 min. For biofilm staining, 200 µl of crystal violet (80 µg/ml in distilled water) was added to the wells, followed by incubation at RT for 5 min. The wells were then washed twice with 1 ml distilled water and dried at 40 • C for 30 min. Biofilm quantification was performed by A 600 determination (microplate reader, Tecan Infinite 200 and Tecan Spark). To account for the different distributions of biofilms within a single well, measurements were performed 100 times within one well, and the average was calculated. If required, a sub-inhibitory concentration of vancomycin (0.78 µg/ml) or 5 µg/ml of the anti-biofilm peptide DJK-5 (De La Fuente-Nunez et al., 2015) were added. To test the biofilm eradication capacity, preformed biofilms (37 • C, 8 h) were incubated for an additional 16 h in the presence of vancomycin at concentrations ranging from 1 to 100 µg/ml. Biofilm staining and CFU determination were performed as described above.

(p)ppGpp Synthetases Show no Significant Effect on Biofilm Formation Under Non-stress Conditions
Recently, it was proposed that the stringent response facilitates biofilm formation in several pathogens, including S. aureus (De La Fuente-Nunez et al., 2014). However, basal medium 2 (BM2) used in this study allowed hardly any S. aureus growth, resulting in a final OD 600 below 1 after overnight growth. Therefore, we first analyzed the impact of (p)ppGpp on biofilm formation using different media. These media included tryptic soy broth (TSB) with the addition of 0.5% glucose and 3% NaCl, which is widely used for S. aureus biofilm analyses (Lade et al., 2019), and CDM (Pohl et al., 2009), used to define the stringent response phenotype in S. aureus. For discrimination between the Reland RelP/RelQ-mediated stringent response, a rel syn mutant (mutation in the synthetase domain, leaving hydrolase activity unaltered), a relP, relQ double mutant and a (p)ppGpp 0 mutant in which all three synthetases were non-functional were included in the analysis. Independent of the (p)ppGpp synthetases, all strains showed the strongest biofilm formation in CDM. Interestingly, in the prototypic biofilm medium TSB, biofilm formation was lower than in CDM ( Figure 1A). However, no significant difference in biofilm formation was observed between the different strains in either medium. In BM2, a trend toward a slightly higher biofilm-forming capacity in the wild type compared to the mutant strains was observed, indicating that under these nutrient-limited conditions, the stringent response might be slightly activated. Thus, (p)ppGpp synthetases have no or little influence on the biofilm formation ability under non-stressed conditions.

RelP and RelQ Regulate Biofilm Formation Under Cell Wall Stress
Subinhibitory concentrations of vancomycin were shown to transcriptionally activate relP and relQ (Geiger et al., 2014). We thus speculated that the vancomycin-induced stringent response may interfere with biofilm formation. The minimal inhibitory concentration (MIC) (1 µg/ml) of vancomycin in planktonically grown bacteria did not differ between the analyzed strains. At a subinhibitory concentration of vancomycin (0.78 µg/ml), the relPQ double mutant and the (p)ppGpp 0 mutant showed significantly reduced biofilm formation compared to the wild type and the rel syn mutant (Figures 1B,C). Under vancomycin treatment, the wild type formed an almost uniform thick layer of biofilm. In contrast, the relPQ mutant and the (p)ppGpp 0 strain showed significantly decreased biofilm formation, with some cell aggregates remaining after the washing procedure (Figures 1B,C). The bacterial survival of the tested strains was not impaired by the subinhibitory concentrations of vancomycin ( Figure 1D). Thus, the RelP-and/or RelQ-dependent changes in biofilm formation were not due to growth inhibition or bacterial killing.

RelP and RelQ Synergistically Affect Biofilm Formation
To determine which of the synthetases impacts biofilm formation under vancomycin conditions, single relP and relQ mutants were analyzed. Both relP and relQ contributed to biofilm formation in the presence of vancomycin. They act synergistically, since the relPQ double mutant showed the lowest biofilm formation (Figure 2). The relPQ double mutant phenotype could be complemented by the integration of either relP or relQ into the chromosome (Figure 2). Thus, the vancomycin-dependent induction of either relP or relQ is sufficient to sustain S. aureus biofilm formation under cell wall stress conditions.

The Biofilm Composition Is Not Affected by the Stringent Response
The biofilm matrix is composed of PIA, proteins or eDNA (O'Gara, 2007;Paharik and Horswill, 2016). We analyzed which matrix components were involved in the observed RelP/Qdependent biofilm alterations under vancomycin treatment. Preformed biofilms were treated with sodium periodate, proteinase K or DNase to selectively degrade PIA, proteins or eDNA matrix components, respectively (Seidl et al., 2008). Without vancomycin, the biofilms formed by the wild type or (p)ppGpp 0 stain were almost completely degraded by proteinase K and DNase treatment, whereas sodium periodate had no effect on the biofilm matrix (Figures 3A,B). To ensure that vancomycin does not impact the biofilm composition, we additionally examined the matrix components after vancomycin treatment. Again, the biofilms consisted almost exclusively of proteins, and eDNA and sodium periodate treatment did not degrade the biofilm matrix. Thus, under the conditions applied here, HG001 forms ica-independent biofilms, and the biofilm composition is not altered by the stringent response.

The Stringent Response Induces Biofilm Formation Independent of Agr and CodY
The quorum-sensing system Agr (especially the target genes psm) (Otto, 2018) and the transcriptional regulator CodY (Stenz et al., 2011) have been identified as key controllers of biofilm structure and detachment. (p)ppGpp synthesis results in the derepression of the CodY regulon and upregulation of Agrdependent psm genes (Geiger et al., 2012). Thus, we hypothesized that Agr and/or CodY activity could interfere with the observed (p)ppGpp-dependent biofilm. However, the mutation of agr or codY did not impact biofilm formation (Figure 4). Thus, under our assay conditions, biofilm formation occurs independent of CodY or Agr. Strain specific effects of Agr (Yarwood et al., 2004) or CodY (Stenz et al., 2011) on biofilm formation were described previously.

(p)ppGpp Contributes to Biofilm Antibiotic Tolerance
Biofilms are normally more tolerant to high concentrations of antibiotics than planktonic cultures. We hypothesized that the stringent response contributes to biofilm antibiotic tolerance in S. aureus. Therefore, biofilm antibiotic tolerance was compared between the wild type and the isogenic (p)ppGpp 0 mutant. Preformed biofilms were exposed to increasing concentrations of vancomycin for 16 h. At the MIC (1 µg/ml for planktonically grown bacteria), vancomycin did not result in biofilm dispersal. However, at concentrations 10-and 100-fold higher than the MIC, the biofilm produced by the (p)ppGpp 0 strain was significantly reduced, whereas the biofilm produced by the wild type was more resistant to vancomycin treatment (Figures 5A,B). Thus, (p)ppGpp contributes to biofilm-related antibiotic tolerance.

The Anti-Biofilm Peptide DJK-5 Exerts Its Effects Independent of (p)ppGpp
Recently, the peptides IDR-1018 (De La Fuente-Nunez et al., 2014) and DJK-5 (De La Fuente-Nunez et al., 2015) were proposed to prevent biofilms due to the specific targeting of intracellular (p)ppGpp. If correct, the peptides are expected to inhibit biofilm formation under stringent conditions in the wild type but not in the pppGpp 0 background. We confirmed that DJK-5 interferes with biofilm formation in S. aureus (Figures 6A,B). However, without vancomycin treatment, biofilm formation by the wild type and pppGpp 0 strains was equally affected by DJK-5, indicating that the effect was independent of (p)ppGpp. The combination of a subinhibitory vancomycin concentration and DJK-5 resulted in the complete inhibition of biofilm formation in the pppGpp 0 strain. This can be explained by bacterial killing of the pppGpp 0 strain through the synergistic action of vancomycin and DJK-5 ( Figure 6C). Thus, (p)ppGpp synthesis in the wild type obviously protects the strain from the action of DJK-5. These findings are in contrast to the assumption that the biofilm-inhibiting activity of DJK-5 is exerted via (p)ppGpp inhibition, as proposed by De La Fuente-Nunez et al. (2014.

DISCUSSION
Here, we show that the small alarmone synthetases RelP and RelQ maintain the biofilm-forming capacity of S. aureus when exposed to subinhibitory concentrations of vancomycin. Both enzymes are part of the cell wall stress regulon and are transcriptionally induced by vancomycin (Geiger et al., 2014). They act synergistically, and the relPQ double mutant can be complemented via the chromosomal integration of either relP or relQ. Thus, the (p)ppGpp synthesis expected to occur upon vancomycin treatment supports biofilm growth, whereas without (p)ppGpp, no biofilms are formed in the presence of vancomycin. How (p)ppGpp promotes biofilm formation remains to be elucidated. (p)ppGpp results in an immediate decrease in intracellular GTP and derepression of the CodY regulon . When CodY is loaded with GTP and/or branched-chain amino acids, it represses many metabolism-related genes, the Agr system and ica gene expression (Majerczyk et al., 2008;Pohl et al., 2009). The impact of CodY on biofilm formation is probably multifactorial and strain dependent FIGURE 6 | DJK-5 interferes with biofilm formation under relaxed and stringent conditions. (A) Biofilms grown with or without vancomycin (0.78 µg/ml) and/or the anti-biofilm peptide DJK-5 (5 µg/ml) in CDM for 24 h. (B) Representative plate stained with crystal violet. (C) CFU of planktonic and biofilm bacteria determined in parallel after static growth for 24 h. Three separate experiments were performed with biological triplicates each. Error bars represent the standard deviation, statistical significance based on ordinary one-way ANOVA (ns: not significant, *: P < 0.1, ****: P < 0.0001). (Stenz et al., 2011;Atwood et al., 2015). codY mutants have also been reported to aggregate, which can be linked with the interaction of PIA and eDNA on the bacterial surface (Mlynek et al., 2020). However, we can exclude the involvement of CodY regulation in the observed biofilm maintenance, since biofilm formation was not altered in a codY-negative background. Additionally, the Agr quorum-sensing system and, thus PSMs (which are strongly dependent on Agr activity), were excluded as mediators of the biofilm phenotype. Thus, the main mechanism for biofilm formation under vancomycin stress remains to be elucidated. (p)ppGpp dependent DNA-release by any of the lytic processes (e.g., autolysins, phages) likely contributes to biofilm formation. Recently, it was shown that mupirocin, a strong inducer of Rel-dependent (p)ppGpp synthesis, causes increased biofilm formation (Jin et al., 2020). Similar to our results, the mupirocin-induced biofilm forms independent of PIA and PSMs and is largely composed of eDNA. Thus, it is likely that the observed biofilm-inducing phenotype induced by mupirocin is similar to the vancomycin-dependent biofilm observed in our study. Jin et al. (2020) found that mupirocin upregulates cidA, encoding a holin-like protein, and that a cidA mutant shows reduced eDNA release. Thus, one may speculate that under our assay conditions, the (p)ppGpp-mediated activation of cidA may also contribute to (p)ppGpp-promoted biofilm formation.
The subinhibitory concentration of vancomycin applied in our standard biofilm assay did not affect bacterial viability, and the MIC in planktonically grown strains did not differ between the analyzed strains. When vancomycin was added to preformed biofilms, the biofilms were protected even at up to a concentration 100-fold higher than the MIC. It has been proposed that antibiotic tolerance and persister formation share common characteristics such as a slow-or non-growing phenotype (Waters et al., 2016). Here, we showed that biofilm tolerance is at least partly (p)ppGpp dependent, since biofilms of the pppGpp 0 strain were significantly better resolved in the presence of high vancomycin concentrations. This seems to contrast with recent results indicating that (p)ppGpp is not involved in persister formation in S. aureus . However, the persister assays were performed under relaxed conditions, and thus, the role of (p)ppGpp might have been missed. Nevertheless, (p)ppGpp synthesis was previously shown to contribute to antibiotic tolerance in S. aureus (Geiger et al., 2014;Bryson et al., 2020) and other pathogens (Nguyen et al., 2011;Bernier et al., 2013). Nguyen et al. (2011) suggested that in Pseudomonas aeruginosa, the stringent response contributes to antimicrobial tolerance in biofilms by reducing oxidative stress. We recently showed that (p)ppGpp in S. aureus activates ROS-detoxifying systems (Horvatek et al., 2020), which might contribute to protection against vancomycin.
Due to the role of (p)ppGpp in biofilm formation and antibiotic tolerance, the (p)ppGpp synthesis pathway is thought to be a promising antimicrobial target. Anti-biofilm peptides have been reported to exert their activity via their ability to reduce (p)ppGpp levels in live bacterial cells (De La Fuente-Nunez et al., 2014. A direct mechanism of action involving the binding of (p)ppGpp and promotion of its intracellular degradation was suggested (De La Fuente-Nunez et al., 2015). We confirmed the biofilm-dissolving effect of DJK-5. However, this was clearly not due to the proposed interaction of the peptides with (p)ppGpp because an even stronger inhibitory effect of DJK-5 was observed in the (p)ppGpp 0 mutant. Interestingly, treatment with DJK-5 and a subinhibitory vancomycin concentration resulted in significantly higher bacterial killing activity and biofilm inhibition in the (p)ppGpp 0 mutant than the wild type. Thus, (p)ppGpp protects against bactericidal DJK-5 activity. These findings are in good agreement with a recent re-analysis of the proposed antibiofilm peptide IDR-1018 in E. coli and P. aeruginosa (Andresen et al., 2016). Genetic disruption of the relA and spoT genes responsible for (p)ppGpp synthesis moderately sensitizes E. coli to IDR-1018, rather than protecting the bacterium (Andresen et al., 2016). While the IDR-1018 and DJK-5 peptides are potent antimicrobials, they do not specifically disrupt biofilms via a direct and specific interaction with the intracellular messenger nucleotide (p)ppGpp. Their alternative mode of action remains to be elucidated.

CONCLUSION
In conclusion, in S. aureus, (p)ppGpp supports biofilm formation under cell wall stress conditions and increases tolerance against vancomycin and the anti-biofilm peptide DJK-5.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.