Identification of the agr Peptide of Listeria monocytogenes

Listeria monocytogenes (Lm) is an important food-borne human pathogen that is able to strive under a wide range of environmental conditions. Its accessory gene regulator (agr) system was shown to impact on biofilm formation and virulence and has been proposed as one of the regulatory mechanisms involved in adaptation to these changing environments. The Lm agr operon is homologous to the Staphylococcus aureus system, which includes an agrD-encoded autoinducing peptide that stimulates expression of the agr genes via the AgrCA two-component system and is required for regulation of target genes. The aim of the present study was to identify the native autoinducing peptide (AIP) of Lm using a luciferase reporter system in wildtype and agrD deficient strains, rational design of synthetic peptides and mass spectrometry. Upon deletion of agrD, luciferase reporter activity driven by the PII promoter of the agr operon was completely abolished and this defect was restored by co-cultivation of the agrD-negative reporter strain with a producer strain. Based on the sequence and structures of known AIPs of other organisms, a set of potential Lm AIPs was designed and tested for PII-activation. This led to the identification of a cyclic pentapeptide that was able to induce PII-driven luciferase reporter activity and restore defective invasion of the agrD deletion mutant into Caco-2 cells. Analysis of supernatants of a recombinant Escherichia coli strain expressing AgrBD identified a peptide identical in mass and charge to the cyclic pentapeptide. The Lm agr system is specific for this pentapeptide since the AIP of Lactobacillus plantarum, which also is a pentapeptide yet with different amino acid sequence, did not induce PII activity. In summary, the presented results provide further evidence for the hypothesis that the agrD gene of Lm encodes a secreted AIP responsible for autoregulation of the agr system of Lm. Additionally, the structure of the native Lm AIP was identified.


INTRODUCTION
The Gram-positive bacterium Listeria monocytogenes (Lm) is an opportunistic, intracellular pathogen that may cause severe, food-borne infections in high-risk groups such as immunocompromised persons, elderly people and pregnant women (Freitag et al., 2009). Lm is able to survive and replicate in a wide range of environments including soil, various food products, and different niches inside its human host (Freitag et al., 2009;Vivant et al., 2013;Ferreira et al., 2014;Gahan and Hill, 2014). In order to adapt to these changing conditions, L. monocytogenes possesses 15 complete two-component systems (Williams et al., 2005) and a number of regulatory circuits (Guariglia-Oropeza et al., 2014). The accessory gene regulator (agr) locus encodes one of these systems and has been shown to be involved in biofilm formation, virulence and survival in the environment (Autret et al., 2003;Rieu et al., 2007;Riedel et al., 2009;Vivant et al., 2015). The prototype agr system was described for S. aureus and consists of the four gene operon agrBDCA (Novick and Geisinger, 2008). Of the four proteins encoded by the agr operon, AgrB is a membrane-bound peptidase that cleaves and processes the agrD-derived propeptide at the C-terminus, catalyzes formation of a thiolactone ring with a central cysteine, and, in combination with the signal peptidase SpsB, effects export and release of the active autoinducing peptide (AIP). Upon accumulation in the extracellular space, this AIP activates a two-component system consisting of AgrC (receptor-histidine kinase) and AgrA (response regulator). Expression of the operon is driven by the P II promoter upstream of agrB and is subject to autoregulation via AgrA. Target genes of the staphylococcal agr system are either directly regulated by AgrA or by a regulatory RNAIII transcribed in the opposite direction from the P III promoter adjacent to P II (Thoendel et al., 2011).
Homologous agr systems have been identified in a number of Gram-positive microorganisms including streptococci, clostridia, lactobacilli, Bacillus sp., and Enterococcus faecalis (Wuster and Babu, 2008). The effects of agr regulation are pleiotropic. In S. aureus, the agr system regulates a wide range of genes involved in biofilm formation, virulence, and immune evasion (Queck et al., 2008;Thoendel et al., 2011). The agr system of Lactobacillus plantarum is involved in regulation of cell morphology and adhesion to glass surfaces (Sturme et al., 2005;Fujii et al., 2008). Similar to the staphylococcal system, the agr-like fsr system of E. faecalis and the agr system of Lm are involved in regulation of biofilm formation and virulence (Autret et al., 2003;Rieu et al., 2007;Riedel et al., 2009;Cook and Federle, 2014). Moreover, in Lm more than 650 genes are directly or indirectly regulated by the agr system as shown by transcriptional profiling of an agrD deletion mutant (Riedel et al., 2009). This suggests that agr systems represent rather global regulatory mechanisms.
Despite similarities on protein level, genetic organization, and phenotypic traits regulated, known agr systems differ regarding their mechanisms of target gene regulation. While in staphylococci, a significant number of agr-dependent genes are regulated by RNAIII (Thoendel et al., 2011), no information on RNAIII transcripts are available in other organisms. In E. faecalis and Lm, the genetic information upstream of the agr operon differs from that of staphylococci in that the preceding gene is transcribed in the same direction as the agr genes and no putative P III promoters have been identified (Qin et al., 2001;Autret et al., 2003). Moreover, despite extensive bioinformatic approaches or transcriptional profiling a regulatory RNAIII has not been identified in Lm (Mandin et al., 2007;Toledo-Arana et al., 2009;Mellin and Cossart, 2012;Wurtzel et al., 2012). This suggests that in Lm (and E. faecalis) target genes are regulated by AgrA and/or other transcriptional regulators affected by AgrAdependent regulation. However, it can not be excluded that the AIP signals through other two-component system besides AgrCA.
Structural information of AIPs is available only for a limited number of species. In S. aureus, four agr specificity groups with different AIPs varying in size from 7 to 9 amino acids (aa) are known (Novick and Geisinger, 2008). Similarly, three agr specificity groups exist in S. epidermidis with AIPs of 8-12 aa (Otto et al., 1998;Olson et al., 2014). The AIP of S. intermedius and S. lugdunensis are 9 and 7 aa in size, respectively (Ji et al., 1997;Kalkum et al., 2003). Outside the genus Staphylococcus, AIPs have been characterized for E. faecalis (11 aa), L. plantarum (5 aa), and C. acetobutylicum (6 aa) (Nakayama et al., 2001;Sturme et al., 2005;Steiner et al., 2012). Most of the known AIPs contain a thiolactone ring formed by the 5 C-terminal aa. Exceptions are the AIPs of C. acetobutylicum and E. faecalis, which have ring structures consisting of 6 and 9 aa, respectively (Nakayama et al., 2001;Steiner et al., 2012). Another common feature is a central cysteine, which is replaced by a serine in some cases, required for thiolactone ring formation.
For staphylococci, E. faecalis and Lm, a contribution of the agr system to virulence gene regulation has been demonstrated and agr-deficient mutants are attenuated (Riedel et al., 2009;Thoendel et al., 2011;Cook and Federle, 2014). Consequently, interference with agr signaling was proposed as a therapeutic approach (Gray et al., 2013). Of note, the specificity of the interaction between the AIP and its cognate receptor AgrA has been used to device improved strategies by fusing the AIP to a bacteriocin to induce lysis of the targeted bacteria (Qiu et al., 2003). The structure of the native AIP of Lm has not been elucidated so far. With the present study, we aim closing this gap in order to further elucidate the components and mechanisms of the agr autoregulatory circuit of Lm and to facilitate future studies on strategies to interfere with cell-cell communication of this important human pathogen.

Bacterial Strains and Culture Conditions
All strains and plasmids used in this study are listed in Table 1. L. monocytogenes was generally incubated in Brain Heart Infusion broth (BHI, Oxoid Ltd) at 30 • C. E. coli strains were grown in lysogeny broth (LB). For solid media, 15 g/l agar were added to the broth before autoclaving. Antibiotics were added if necessary. Where appropriate, kanamycin was used at a final concentration of 50 (for E. coli strains) and 15 µg/ml chloramphenicol were used for both species. For Lm strains carrying a chromosomal copy of pPL2 derivatives chloramphenicol was used at 7 µg/ml.

Generation of Recombinant Strains
Primers used for cloning or sequencing purposes are listed in Table 2. To study transcriptional activity of the agr operon, the P II promoter upstream of agrB (Rieu et al., 2007) was amplified with Phusion R polymerase (Thermo Fisher Scientific) using primers PII_fwd_SalI and PII_rev and chromosomal DNA of Lm EGD-e wildtype (WT) as template. The obtained PCR fragment was digested with SalI and cloned in frame in front of the luciferase reporter into SalI/SwaI-cut pPL2lux (Bron et al., 2006). The ligation mix was transformed into E. coli ElectroMax TM DH10B (Thermo Fisher Scientific), and the resulting plasmid pPL2luxP II was verified by restriction analysis and amplification of the cloned P II promoter using primers PII_fwd_SalI and luxA_rev with subsequent Sanger sequencing of the PCR fragment by a commercial service provider (Eurofins, Germany). The plasmid was transformed into electrocompetent Lm EGD-e WT or agrD (Riedel et al., 2009) as described previously (Monk et al., 2008) creating Lm EGD-e::pPL2luxP II and agrD::pPL2luxP II . In both strains, successful chromosomal integration of pPL2luxP II at the correct site (tRNA Arg ) was verified using primers PL95 and PL102 (Lauer et al., 2002).
For homologous overexpression of agrBD, a PCR fragment containing both genes was amplified using primers NZagrBD_fwd and NZagrBD_rev and chromosomal DNA of Lm EGD-e as template. The PCR product was digested with NcoI and SacII and ligated as exact transcriptional fusion to the constitutive P 44 promoter into NcoI/SacII digested pNZ44 (McGrath et al., 2001) to yield pNZ44agrBD. The product was transformed into E. coli DH10B. Clones were screened for plasmid containing the correct insert by PCR using primers NZ-confirm_fwd and NZ_colony_rev and sequencing of the PCR product. The correct plasmid as well as the empty vector (pNZ44) were transformed in electrocompetent Lm agrD generated as described previously (Monk et al., 2008).
For heterologous AIP production, agrBD or agrB alone were amplified using primer pairs agrBD_NdeI_fwd/ agrBD_BamHI_rev and chromosomal DNA of Lm EGD-e WT or agrD. Following restriction with NdeI and BamHI both PCR products were ligated into NdeI/BamHI digested pET29a(+) (Merck Millipore). This fuses the PCR products to the T7 promoter creating pET29a_agrB and pET29a_agrBD,

AIP Production in E. coli
For heterologous AIP production, pET29a_agrB or pET29a_agrBD were transformed into E. coli BL21(DE3) (New England Biolabs) and transformants were selected on LB agar containing kanamycin. Four single colonies were streaked onto two LB agar plates containing kanamycin with or without 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG). A clone showing good growth in the absence of IPTG but reduced growth in its presence was selected and a single colony was inoculated into 5 ml of LB medium and grown o/N on a rotary shaker at 37 • C. Using the o/N culture, 500 ml LeMaster and Richards minimal medium (Paliy and Gunasekera, 2007) containing 50 mM glucose were inoculated to a final OD 600 of 0.1 and incubated on a rotary shaker at 37 • C to an OD 600 of 0.8. At this stage, expression was induced by addition of 1 mM ITPG. Following incubation under the same conditions for an additional 2 h, bacterial cells were pelleted via centrifugation (3000 × g, 30 min and 4 • C) and supernatants were collected, filter sterilized, frozen in liquid nitrogen and lyophilized. Lyophilized samples were stored at −20 • C until further analysis by LC-MS/MS.   with ESI Jet Stream Technology using the following conditions: drying gas flow rate of 10 l/min with a gas temperature of 250 • C, nebulizer with 40 lb per square inch gauge, sheath gas flow rate of 10 l/min, sheath gas temperature of 300 • C, capillary voltage of 4000 V, and fragmentor voltage of 170 V. The collision energy was set by formula with 4.5 slope and 10 offset. Data analysis was performed using Mass Hunter Workstation Software (Ver.B.05.519.0, Agilent Technologies) and the "Find compounds by formula" algorithms. Synthetic peptides were analyzed using the same conditions as the recombinant peptides expressed in E. coli to compare retention time, accurate mass and fragmentation patterns.

Invasion Assay
Invasion of Lm into Caco-2 cells was tested using a standard gentamycin protection assay essentially as described previously (Riedel et al., 2009). Briefly, Caco-2 cells were cultured in DMEM supplemented with 10% (v/v) fetal calf serum (FCS), 10 mM L-glutamine, 1% (v/v) penicillin/streptomycin and 1% (v/v) non-essential amino acids (NEAA) at 37 • C and a 5% CO 2 atmosphere. Cells were seeded to a density of 2 × 10 5 cells per well in a 24 well plate and cultivated to a monolayer for 4 days. One day prior to the experiment, culture media without antibiotics was added. A fresh o/N culture of the indicated bacterial strains was diluted 1:10 in 10 ml fresh BHI and grown to mid-exponential phase (OD 600 = 0.8). Where appropriate, peptide R5T0 was added (5 µM final concentration). Bacteria were pelleted and diluted in DMEM containing 10 mM Lglutamine and 1% NEAA to 10 8 colony forming units per ml (cfu/ml) (OD 600 = 0.5). 1 ml of this suspension was added to Caco-2 cells in quadruplicates (MOI = 100). Cells were incubated for 1 h to allow invasion of bacteria. To kill remaining extracellular bacteria, cells were washed once with PBS and 1 ml DMEM containing 10 µg/ml gentamicin (Gibco R ) was added to the cells. After 1 h of incubation, cells were washed twice with PBS, lysed with ice-cold water and cfu/ml were determined by plating serial dilutions on BHI agar.

Statistical Analysis
All experiments were conducted in at least three biological replicates. Results were analyzed by Student's t-test or ANOVA with Bonferroni post-test analysis to correct for multiple comparisons using GraphPad Prism (version 6) as indicated in figure legends and Supplementary Data Sheet 1. Differences between different strains or conditions were considered statistically significant at p < 0.05.

RESULTS
P II -Activity in Lm EGD-e P II promoter activity was analyzed in Lm EGD-e::pPL2luxP II and agrD::pPL2luxP II during growth in BHI medium at 30 • C ( Figure 1A). No differences in growth or final OD 600 were observed between the two strains ruling out an effect of growth on luciferase activity. In the WT background, a significant increase in P II -dependent luciferase activity was observed during exponential growth with a peak in late exponential phase. By contrast, no luminescence above background could be detected for the agrD-deficient strain throughout the experiment. This suggests that the AIP is required for transcriptional activity of P II . AIPs are usually secreted into the extracellular environment. In order to confirm that the AIP of Lm is acting as an extracellular peptide, similar growth experiments were conducted using co-incubation of AIP producer and reporter strains in different combinations ( Figure 1B). As expected, the agrDdeficient reporter strain showed no P II activity when incubated with Lm EGD-e agrD. However, high levels of luminescence were observed using the same reporter strain in combination with Lm agrD pNZ44agrBD, a agrD derivative expressing agrBD from the P44 promoter on pNZ44. Luminescence in this setup was significantly higher compared to co-cultures of the WT reporter with the agrD deletion mutant or the agrD-deficient reporter strain with Lm EGD-e pNZ44 (i.e., the empty vector control) suggesting that AIP levels produced by Lm agrD pNZ44agrBD are higher than that of the WT.

P II Activation by Synthetic AIP Candidates
Upon several attempts we were unable to identify the active AIP in supernatants of Lm EGD-e WT or the AIP overproducing strain agrD pNZ44agrBD grown in either BHI or modified Welshimer's broth. Sequence alignment of AIPs with a resolved structure, revealed that most AIPs consist of a 5 aa thiolactone ring with N-terminal tail varying from 0 to 7 aa (Figure 2A). Using this information, a range of peptides based on the AgrD sequence of Lm EGD-e were synthesized consisting of a thiolactone ring of 4-6 aa and an N-terminal tail of 0-5 aa ( Figure 2B). The effect of these peptides on P II -driven luciferase activity was tested using the reporter strains Lm EGDe::pPL2luxP II and agrD::pPL2luxP II . At 5 µM, none of the peptides had a measurable effect on growth of the reporter strains (Supplementary Figures S1A,B). The peptide R5T0 consisting of a 5 aa thiolactone ring with no N-terminal tail slightly increased P II -driven luminescence in the WT reporter strain during the first 4 h of the experiment (Figure 3A). However, at later stages luminescence was comparable to the control, i.e., reporter without peptide. Interestingly, some of the tested peptides (R5T1, R5T2, R5T4, and R5T5) significantly inhibited luminescence of the WT reporter strain. More importantly, some of the peptides (R5T0, R5T1, R5T4, and R5T5) induced luminescence by the agrD reporter strain (Figure 3B). The most potent inducer of P II activity was the peptide R5T0, i.e., a cyclic pentapeptide with the amino acid sequence Cys-Phe-Met-Phe-Val (CFMFV). At concentration of 5 and 50 µM, R5T0 also induced luminescence above control levels during the first 4 h in the WT reporter ( Figure 3C) and for up to 7 h in the agrD-deficient reporter ( Figure 3D). This suggests that the most likely candidate for the native AIP of Lm EGD-e is the peptide R5T0.

The Synthetic AIP Restores the Invasion Defect of Lm agrD
Deletion of agrD and thus lack of a functional AIP results in reduced promoter activity of virulence factors and attenuated virulence (Riedel et al., 2009). In order to check if R5T0 is not only able to induce P II activity but also functionally complement the agrD mutant, invasion assays were performed with Lm EGD-e agrD grown in the presence and absence of R5T0 (Figure 4). As observed previously, deletion of agrD results in reduced invasion into Caco-2 intestinal epithelial cells and this defect was genetically complemented by integration of pIMK2agrD, i.e., a plasmid for constitutive expression of agrD (Riedel et al., 2009). More importantly, growth in the presence of 5 µM R5T0 completely restored invasion of Lm EGD-e agrD to WT levels.

Heterologous Production of the Lm AIP in E. coli
In a further approach to identify the AIP of Lm, the agrBD genes were expressed in E. coli using the IPTG-inducible pET29a system. Using LC-MS, a prominent signal was identified in supernatants of an induced culture of E. coli BL21 pET29a_agrBD ( Figure 5A) with a mass of 627.2549 (Figure 5B). This signal was absent in the non-induced culture or supernatant of a control strain only expressing agrB (Supplementary Figure S2). In order to confirm the identity of the overexpressed peptide, analysis of the P II -activating synthetic peptide R5T0 was performed. Interestingly, the chromatogram of R5T0 yielded two peaks in close vicinity (Figure 5A). Both peaks correspond to peptides with identical mass and fragmentation pattern ( Figure 5B). However, the different retention times and peak areas indicate that the two peaks represent stereoisomers or conformational isomers at different concentrations.
The peptide present in the supernatant of E. coli BL21 pET29a_agrBD and both peaks of R5T0 had almost identical global masses ( Figure 5B). Moreover, all three peptides showed highly similar fragmentation patterns ( Figure 5B) and several signals of the MS/MS spectra correspond to fragments of R5T0 at a mass accuracy better than 2 ppm ( Table 3; for corresponding structures see Figures 5C,D). These results clearly indicate that the listerial AIP is a cyclic pentapeptide with the amino acid sequence CFMFV forming a thiolactone ring, i.e., the structure of the synthetic peptide R5T0.

Specificity of the Lm AIP
Known AIPs differ greatly in sequence, length and structure among species and even strains (Figure 2A) and different AIPs of S. aureus display cross-inhibition (Ji et al., 1997). Similar to the AIP of Lm, the AIP of L. plantarum is a cyclic pentapeptide yet with a different sequence (Sturme et al., 2005). Further experiments were performed to test if P II activation is specific for the Lm AIP or if the L. plantarum AIP is also able to activate P II (Figure 6). As observed in the previous experiments, R5T0 slightly enhanced P II -driven luciferase activity in Lm EGD-e::pPL2luxP II ( Figure 6A) and was a potent inducer of P II activity in the AIP-negative reporter strain Lm EGD-e agrD::pPL2luxP II . By contrast, in both reporter strains the L. plantarum AIP had no effect on P II activity.

DISCUSSION
Signaling peptides, also referred to as AIPs, are produced by a wide range of Gram-positive microorganisms (Wuster and Babu, 2008) and serve various purposes (Thoendel and Horswill, 2010). The best studied AIP system is the agr locus of S. aureus and homologous systems have been identified in a variety of Gram-positives (Wuster and Babu, 2008). In S. aureus, the agr system is a rather global regulatory circuit affecting a large number of genes and different phenotypic traits (Thoendel et al., 2011). Similarly, deletion of agrD in Lm affects more than 600 genes and phenotypically affects biofilm formation and virulence in vitro and in vivo (Rieu et al., 2007;Riedel  et al., 2009). However, while absence of agr signaling is linked with enhanced biofilm formation of S. aureus (Vuong et al., 2000), agr mutants of Lm display reduced biofilm formation under the conditions monitored (Rieu et al., 2007;Riedel et al., 2009).
Previous studies have already indicated that, like the staphylococcal system, the Lm agr locus is subject to positive autoregulation involving a diffusible factor, probably the agrDencoded AIP involved in regulation. Transcription levels of the agr operon were greatly reduced in agr-deficient Lm mutants (Rieu et al., 2007;Riedel et al., 2009;Garmyn et al., 2012). Also, the biofilm defect of a agrD mutant was complemented when bacteria were grown in the reconstituted culture supernatants of the WT or in the presence of small amounts of WT bacteria (Riedel et al., 2009). The presented results further strengthen the hypothesis that agrD encodes a secreted AIP that positively regulates the agr system of Lm.
In the agrD mutant, no activity of the agr promoter could be observed ( Figure 1A) and promoter activity was restored when the agrD-deficient reporter strain was co-cultured with a strain carrying a plasmid for constitutive expression of agrBD ( Figure 1B).
The presented results provide further evidence that, in Lm, agrD actually encodes the propeptide, which is processed released into the extracellular environment where it acts as an AIP. Moreover, our data suggests that the native AIP is a cyclic pentapeptide R5T0 consisting of the amino acids (from N-to C-terminus) Cys, Phe, Met, Phe, Val. A peptide with this structure was found in the culture supernatant of a recombinant E. coli strain expressing AgrBD ( Figure 5) and a synthetic peptide with identical structure was able to potently induce activity of the P II promoter of the agr system (Figure 3) and to functionally complement the invasion defect in a agrD mutant (Figure 4).
Induction of luciferase activity in the agrD reporter upon co-cultivation with the AIP producing WT strain ( Figure 1B) indicates that at least some of the AIP must be present in culture supernatants. However, we were unable to identify the native peptide in supernatants of Lm EGD-e grown in complex media (brain heart infusion) or modified Welshimer's broth. This may be explained by the high levels of peptides in brain heart infusion, which makes identification impossible by LC-MS/MS. In modified Welshimer's broth Lm only grows to low final optical densities and thus any secreted peptide will also be present at low concentrations especially when subject to positive autoregulation and fully induced only at high cell densities. Further studies will be needed to quantify actual AIP concentrations in culture supernatants and the threshold required to activate PII and target gene regulation.
Interestingly, four different synthetic peptides with a fivemembered thiolactone ring and varying tail length had inhibitory activity on the agr promoter in the WT reporter strain, which itself is able to produce the native AIP. Since agr mutants of Lm display attenuated virulence (Autret et al., 2003;Riedel et al., 2009), this suggests that these peptides are antagonists of the native AIP and may represent a potential supplementary or alternative therapeutic approach as proposed for S. aureus and other pathogens (Gray et al., 2013). Interestingly, they also exhibited P II activation in the agrD reporter to varying degrees. This may indicate that these peptides compete with R5T0 or the native AIP for binding to the receptor but their affinity and/or activity is lower. Thus, of the four candidate peptides, the best antagonist of the native AIP is probably R5T2, which efficiently blocks P II activity in the WT but activates luminescence inly marginally in the mutant reporter.
A striking difference between the agr systems of S. aureus and Lm is the structural diversity of the AIPs. Within the species S. aureus, four specificity groups of strains with different AIP are found and these groups show cross-inhibition (Novick and Geisinger, 2008). By contrast, the AgrD propeptides of the genus Listeria are rather conserved and the species Lm, L. innocua, L. ivanovii, L. welshimeri, L. seeligeri, and L. marthii have identical (predicted) AIP sequences (Supplementary Figure  S3A) suggesting cross-reactivity. Moreover, phylogenetic analysis based on 16S rRNA gene sequences reveals that Listeria sp. that share identical AIP sequences form a cluster that separates from the other species indicating that they are more closely related (Supplementary Figure S3B). With the exception of C. acetobutylicum, phylogenetic trees calculated using concatenate AgrA, AgrB, AgrC, and AgrD sequences are in line with trees inferred from 16S sequences (Wuster and Babu, 2008). This suggests that agr systems are generally inherited vertically. It has been proposed that C. acetobutylicum, whose AgrD sequence is almost identical to that of Listeriaceae, is the only known case of horizontal transfer of an agr system (Wuster and Babu, 2008). Further experimental data comparing the Lm AIP with the AIP of L. plantarum, which also consist of a five cyclic pentapeptide although with different aa composition, indicates that the Lm agr system is specific for the AIP of those Listeria sp. that share a conserved AgrD sequence but does not respond to the cyclic pentapeptide AIPs of other organisms. This also suggests that intervention strategies based on antagonistic peptides targeting the agr systems of Lm (and other organisms) are specific for organisms with identical AIPs.
In summary, the presented data shows that the agrD of Lm EGD-e encodes a secreted peptide consisting of a fivemembered thiolactone ring, which has autoinducing activity. Moreover, the identification of several synthetic peptides with antagonistic activity proposes a potential option to treat Lm infections or inhibit biofilm formation as suggested by others previously.

AUTHOR CONTRIBUTIONS
CR conceived the study. MZ, MW, and AS-K carried out experiments. MZ, AS-K, BB, and CR analyzed data. MZ, AS-K, BB, and CR drafted the manuscript and all the authors contributed to preparing the final version of the manuscript. All authors read and approved the final manuscript.

FUNDING
The study was conducted by intramural funding of the Universities of Ulm and Stuttgart. The funders had no role in design of the study or analysis and interpretation of the data.