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.

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^® 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 ΔagrD (Riedel et al., 2009) as described previously (Monk et al., 2008) creating Lm EGD-e::pPL2luxP_II and ΔagrD::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 P₄₄ 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 ΔagrD 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 ΔagrD. 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, respectively. Both plasmids were verified for correct cloning by restriction analysis and Sanger sequencing of inserts.

For luciferase reporter assays, growth experiments were performed in white 96-well microtiter plates with transparent bottom (BRANDplates^® pureGrade^TM S). A single colony was inoculated into BHI and grown over night (o/N; i.e., approx. 16 h). Following o/N growth, cultures were diluted to an optical density at 600 nm (OD₆₀₀) of 0.01 in fresh, sterile BHI. For co-cultivation of AIP producer and reporter strains, o/N cultures of both strains were used to inoculate BHI medium to a OD₆₀₀ of 0.01 and then mixed at a 1:1 ratio. 200 μl aliquots of this mix were transferred into individual wells of the microtiter plates (each condition in triplicates). Plates were incubated at 30°C in a Tecan Infinite M200 plate reader and OD₆₀₀ and luminescence intensity were measured every hour.

Synthetic peptides were purchased from Peptide Protein Research Ltd (UK) in lyophilized form with >70% purity. Peptides were reconstituted in dimethyl sulfoxide (DMSO) at 2 mM and stored at -20°C until further use. For experiments, these stocks were diluted as appropriate in 25% (v/v) DMSO in phosphate-buffered saline (PBS) to give the final concentrations as indicated. To test the effect of peptides on P_II activity, reporter strains (Lm EGD-e::pPL2luxP_II or ΔagrD::pPL2luxP_II) were grown o/N and diluted to an OD₆₀₀ of 0.01 in fresh BHI. 180 μl aliquots were distributed in 96 well microtiter plates (each condition in triplicate) and incubated at 30°C for 2 h. At this stage, 20 μl of diluted peptides were added to obtain the indicated final concentrations (5 nM–50 μM) and plates were incubated at 30°C in a Tecan Infinite M200 plate reader with hourly OD₆₀₀ and luminescence intensity measurements.

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 OD₆₀₀ of 0.1 and incubated on a rotary shaker at 37°C to an OD₆₀₀ 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.

The lyophilized supernatants of recombinant E. coli strains were reconstituted in a 25:35:35:5 H₂O:Isopropanol:CH₃CN:HCOOH mixture and diluted 1:10 in H₂O. 5 μl were injected into a reverse-phase column with corresponding guard column (Aeris^TM PEPTIDE 3.6u XB-C18 150 × 2.1 mm, Security Guard^TM ULTRA 2 × 2.1 mm guard column, Phenomenex). A constant flow rate of 0.4 ml/min was applied. Mobile phase A consisted of water with 0.2% (v/v) formic acid and mobile phase B was acetonitrile with 0.2% (v/v) formic acid. Elution program was: isocratic hold at 5% B for 5 min followed by a linear gradient from 5 to 45% B over 80 min. After each sample, the column was washed with 90% B for 10 min and equilibrated at starting conditions. Data was obtained in positive auto MS/MS mode on an Agilent 6540 Accurate-Mass Quadrupole (LC-Q-TOF/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 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% CO₂ atmosphere. Cells were seeded to a density of 2 × 10⁵ 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 (OD₆₀₀ = 0.8). Where appropriate, peptide R5T0 was added (5 μM final concentration). Bacteria were pelleted and diluted in DMEM containing 10 mM L-glutamine and 1% NEAA to 10⁸ colony forming units per ml (cfu/ml) (OD₆₀₀ = 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^®) 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.

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.

P_II promoter activity was analyzed in Lm EGD-e::pPL2luxP_II and ΔagrD::pPL2luxP_II during growth in BHI medium at 30°C (Figure 1A). No differences in growth or final OD₆₀₀ 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 agrD-deficient reporter strain showed no P_II activity when incubated with Lm EGD-e ΔagrD. However, high levels of luminescence were observed using the same reporter strain in combination with Lm ΔagrD pNZ44agrBD, a ΔagrD 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 ΔagrD pNZ44agrBD are higher than that of the WT.

Upon several attempts we were unable to identify the active AIP in supernatants of Lm EGD-e WT or the AIP overproducing strain ΔagrD 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 EGD-e::pPL2luxP_II and ΔagrD::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 ΔagrD 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.

Deletion of ΔagrD 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 ΔagrD mutant, invasion assays were performed with Lm EGD-e ΔagrD 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 ΔagrD to WT levels.

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.

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 ΔagrD::pPL2luxP_II. By contrast, in both reporter strains the L. plantarum AIP had no effect on P_II activity.