Angicin, a novel bacteriocin of Streptococcus anginosus

As a conserved defense mechanism, many bacteria produce antimicrobial peptides, called bacteriocins, which provide a colonization advantage in a multispecies environment. Here the first bacteriocin of Streptococcus anginosus, designated Angicin, is described. S. anginosus is commonly described as a commensal, however it also possesses a high pathogenic potential. Therefore, understanding factors contributing to its host colonization and persistence are important. A radial diffusion assay was used to identify S. anginosus BSU 1211 as a potent bacteriocin producer. By genetic mutagenesis the background of bacteriocin production and the bacteriocin gene itself were identified. Synthetic Angicin shows high activity against closely related streptococci, listeria and vancomycin resistant enterococci. It has a fast mechanism of action and causes a membrane disruption in target cells. Angicin, present in cell free supernatant, is insensitive to changes in temperature from − 70 to 90 °C and pH values from 2 to 10, suggesting that it represents an interesting compound for potential applications in food preservation or clinical settings.

(SilD/E). In S. intermedius the putative bacteriocin production was reported to be induced by the addition of SilCR, but the bacteriocin itself has not been characterized experimentally. However, even though similar genes are present in S. anginosus neither bacteriocin production nor the sil locus has been investigated in this species.
To study the bacteriocin production of S. anginosus, we screened multiple S. anginosus strains for their ability to inhibit closely related bacterial species. Through targeted mutations, the genetic background for bacteriocin production of S. anginosus was identified. The novel bacteriocin was designated as Angicin and phenotypic traits like spectrum of activity, sensitivity towards heat, pH and enzymes as well as the mechanisms of inducing bacteriocin production are elucidated.

Results
Isolation of the bacteriocin producing S. anginosus strain BSU 1211. Producing bacteriocins is a common trait of streptococci. Screening 95 clinical Streptococcus anginosus isolates for strains inhibiting the growth of the closely related streptococcal species Streptococcus anginosus SK 52, Streptococcus constellatus, Streptococcus intermedius and Streptococcus pyogenes (Supplementary Table S1) led to the identification of the strain S. anginosus BSU 1211. Out of the 95 clinical S. anginosus isolates 44 (46%) were able to inhibit at least one of these streptococcal indicator strains, while strain BSU 1211 appears to be a potent bacteriocin producer that was active against multiple species and caused the biggest inhibition zone of all strains. When tested in a onelayer radial diffusion assay (RDA) it shows activity against the S. anginosus type strain SK52 and other streptococci like S. constellatus, S. intermedius and S. pyogenes (Fig. 1). In addition, the pathogen Listeria monocytogenes as well as other listerial species like Listeria ivanovii and Listeria grayi are strongly inhibited by S. anginosus BSU 1211 (Fig. 1). However, for both L. monocytogenes and L. ivanovii sometimes colonies could be seen in the inhibition zones.
Subsequently, the clinically relevant ESKAPE pathogens, consisting of Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii and Pseudomonas aeruginosa were investigated for a susceptibility towards bacteriocins of S. anginosus BSU 1211. These species cause nosocomial infections and are associated with high mortality 27,28 . However, no activity against any of these strains was seen. Genetic background of bacteriocin production. To identify genes responsible for bacteriocin production in S. anginosus BSU 1211 the sil locus and the surrounding regions of the genome were sequenced. In S. intermedius and other species of the SAG, a putative bacteriocin encoding region is adjacent to the quorum sensing locus sil 12 . We found that all sil genes are present in S. anginosus BSU 1211 and that an adjacent region is present, resembling the putative bacteriocin accessory region described in Mendonca et al. (Fig. 2). The open reading frames (ORF) were designated bacteriocin like peptide3 (blp3) in accordance with similar findings in Figure 1. Spectrum of activity of S. anginosus BSU 1211. In a one-layer radial diffusion assay (RDA) S. anginosus BSU 1211 was tested for an effect against several target species. S. anginosus BSU 1211 shows activity against oral streptococci and several listeria species. Depicted is the mean ± standard deviation of at least five experiments.
Regulation and induction of bacteriocin production. Even though the strains BSU 1326, 1370 and 1401 encode the same bacteriocins as S. anginosus BSU 1211 no or only moderate inhibition of target strains can be observed in a one-layer RDA (Fig. 3). To investigate if bacteriocin production in these strains is dependent on the sil locus, synthetic autoinducing peptide SilCR was added in a one-layer RDA. Inhibition zones increase as soon as SilCR SAG-C is administered to the bacterial culture (Fig. 3). Without the addition of SilCR SAG-C S. anginosus BSU 1326 and 1370 are not able to inhibit any target strains. For S. anginosus BSU 1326 inhibition zones with a diameter of 0.36 cm for S. constellatus and 0.47 cm for L. monocytogenes were observed in the presence of 100 ng SilCR SAG-C . Similar effects occurred for the strain BSU 1370 and BSU 1401 (Fig. 3). However, the addition Blp3.1, blp3.4 and blp3.6 are predicted as putative bacteriocins. Blp3.3 and blp3.5 were predicted as putative immunity proteins. SilA, silB, silCR, silD and silE form a quorum sensing system. In red the deleted region of S. anginosus BSU 1211∆blp3 is demonstrated. Schematic overview was constructed with SnapGene. www.nature.com/scientificreports/ of SilCR SAG-C did not lead to larger inhibition zones for S. anginosus BSU1211, suggesting that maximal bacteriocin expression is already present in this strain. It is possible that in strain BSU 1211 bacteriocins are expressed constitutively. The autoinducing peptide SilCR SAG-C alone did not cause any inhibition of target strains.
The blp3 region. Often all genes necessary for bacteriocin production are clustered in one genetic region.
To verify that in S. anginosus BSU 1211 the blp3-region is indeed responsible for the bacteriocin production and thereby the inhibition of target strains, a deletion mutant comprising the entire blp3 region was constructed (S. anginosus BSU 1211∆blp3) (Fig. 2). S. anginosus BSU 1211∆blp3 failed to inhibit any strain that was sensitive towards S. anginosus BSU 1211. The inhibition of these strains could be reestablished by complementing S. anginosus BSU 1211∆blp3 with a plasmid (pAT18-blp3) that encoded the deleted blp3 region. Maximum activity of the complementation mutant against L. monocytogenes, L. grayi or L. ivanovii could be observed with adding synthetic SilCR SAG-C (Fig. 4). As controls, the inhibition of L. monocytogenes, L. grayi, L. ivanovii and S. constellatus by S. anginosus SK52 (a strain which lacks the sil and blp3 region) transformed with pAT18-blp3 and S. anginosus BSU 1211∆blp3 transformed with the empty plasmid was investigated. Neither one of these control strains showed an inhibition of target strains. This data clearly shows that the sil as well as the blp3 region are necessary for bacteriocin production, however the question arises which of the predicted bacteriocins is Depicted is mean ± standard deviation of at least five independent experiments. A significant difference between SilCR SAG-C supplemented and untreated S. anginosus strains was tested with a Mann-Whitney-U-test (***p < 0.001). The effect of completely deleting blp3 or only deleting the predicted bacteriocin genes was tested in a one-layer radial diffusion assay against the four most prominent target strains S. constellatus, L. monocytogenes, L. grayi, and L. ivanovii. Furthermore, a S. anginosus BSU 1211∆blp3 was complemented with the blp3-region encoded on plasmid pAT18. The complementation mutant was supplemented with 100 ng SilCR SAG-C . Shown is the mean ± standard deviation of at least five independent experiments. Significant differences between mutant and wildtype strain were calculated with a Mann-Whitney-U-test (*p < 0.05 www.nature.com/scientificreports/ functional and responsible for the activity or if more than one bacteriocin is produced and active, like it is the case in other streptococci 30,31 or if the observed activity arises from synergistic effects of encoded peptides as described for Garvicin KS of Lactococcus garvieae 32 . To investigate this, single putative bacteriocin genes (blp3.1, blp3.4 and blp3.6) were chromosomally deleted and subsequently these mutant strains were assessed for their ability to inhibit the four most prominent target strains (L. monocytogenes, L. ivanovii, L. grayi, S. constellatus). Deleting blp3.4 caused a complete loss of inhibition activity (Fig. 4). Unfortunately, BSU 1211∆blp3.1 contained an additional point mutation in the non-coding area between blp3.1 and blp3.2. However, it is not suspected that this mutation has any influence on bacteriocin production. A deletion of blp3.1 and blp3.6 did not alter antimicrobial activity. Thus, we conclude that Blp3.4, subsequently labeled Angicin, is the bacteriocin responsible for the antimicrobial activity of S. anginosus BSU 1211. By PCR analysis the blp3.4 gene was found in 41% (38 of 92) S. anginosus strains.
Angicin activity. Angicin is produced as an 84 amino acid long prepeptide. Typically, class II bacteriocins possess a leader peptide, which is cleaved off during export, resulting in the mature and active peptide. The first 30 amino acids of the 84 amino acid long prepeptide form a double glycine leader peptide which after processing leads to the 54 aa long mature Angicin. The peptide sequence of Angicin was compared to similar bacteriocins, like Bovicin 255, Garvicin Q and BacSJ. While Bovicin variant 255 has a sequence identity of approximately 64%, Garvicin Q and BacSJ only have approximately 42% sequence identity ( Supplementary Fig. S1). To verify that Angicin is accountable for the antimicrobial activity of S. anginosus BSU 1211 it was chemically synthesized. Angicin was synthesized without its respective leader peptide and without any post translational modifications. The experimental mass of synthetic Angicin was 6053.1 Da, as determined by mass spectrometry, closely matching the theoretic mass of this 54 amino acid peptide (6052.9 Da). The isoelectric point of the peptide was determined at 10.2 by PSL Heidelberg, while computational analysis of the amino acid sequence with Bachem gave an isoelectric point at 9.6 and a net charge of 4 at neutral pH (https:// www. bachem. com/ servi ce-suppo rt/ pepti de-calcu lator/). The synthetic peptide was dissolved in ultra-pure water, and its activity was examined in a two-layer RDA against L. monocytogenes. Concentrations of 50 μg/ml Angicin caused an inhibition zone of 1.09 ± 0.16 cm, and even concentrations of 1.56 μg/ml Angicin were still able to inhibit growth of L. monocytogenes (Fig. 5a). This value corresponds well to a MIC determination of 3.125 μg/ml against L. monocytogenes and L. grayi (Supplementary Table S2). We then tested as to whether or not Angicin shows the same spectrum of activity as the strain S. anginosus BSU 1211. Based on a two-layer RDA (Fig. 5b) we find that strains of L. monocytogenes, L. grayi, L. ivanovii and S. constellatus, which are susceptible towards S. anginosus BSU1211 (Fig. 1), are also sensitive towards Angicin. However, the inhibition zones of S. constellatus were not completely clear in all experiments. S. anginosus SK52 showed a similar inhibition zone size like S. constellatus. In a second step, we investigated the Angicin susceptibility of a wide range of bacterial species including ESKAPE pathogens. We find that some of these pathogens, such as S. aureus, B. subtilis, are completely insensitive towards Angicin, while others showed a moderate inhibition at Angicin concentrations of 100 μg/ml (E. coli, K. pneumoniae, P. aeruginosa, A. baumannii). Interestingly, among these Angicin-sensitive ESKAPE pathogens was Enterococcus faecium (vancomycin resistant, VRE), which causes nosocomial infections and poses a huge health concern due to rising numbers and limited treatment options 33 . The tested VRE strain was as sensitive in our assay as the listeria species. was prepared and tested for activity. It showed no antimicrobial activity against L. monocytogenes. In a next step, the activity of CFS of BSU 1211 after exposure to heat or pH-changes was investigated. Results are summarized in Table 2 and demonstrate that the S. anginosus CSF can tolerate up to 1 h of 60 °C, an acidic pH of 2, and can be kept at − 20 °C and − 70 °C for one year without loss of activity. However, exposure to 100 °C for 30 min as well as treatment with proteinase K destroyed the bacteriocin activity of CFS. A treatment with lipase had no effect on CFS activity. As a control, CFS of S. anginosus BSU 1211∆blp3 was challenged with the same treatments, but no inhibition of target strains was observed with BSU 1211∆blp3 CFS. We could show that synthetic Angicin resembled the activity pattern of BSU 1211 CFS when exposed to the above-mentioned heat challenges and enzyme treatments. In contrast to BSU 1211 CFS, however, we find that synthetic Angicin caused clear inhibition zones of L. monocytogenes and it showed no reduced activity upon heating at 90 °C for 1 h. To investigate, if the Angicin activity of CFS can also be demonstrated in liquid culture experiments, we measured the growth of L. monocytogenes in the presence of 25% CFS ( Supplementary Fig. S3). We find that L. monocytogenes is significantly inhibited in the presence of 25% CFS from the wildtype strain BSU 1211 (p-value: 0.0006-6 h), while 25% CFS from the deletion mutant S. anginosus BSU 1211∆blp3 slightly enhances growth, when compared to medium without CFS. Furthermore, the activity of Angicin was evaluated in a bacterial survival assay. Since 25% CFS was the best concentration in the inhibition experiment, the same concentration was chosen for a survival assay. In this assay L. monocytogenes, L. grayi and L. ivanovii were investigated. An overnight culture of the indicated strains was inoculated freshly in the morning and grown to an O.D. 600 nm of 0.1. To the same amount of bacteria, resuspended in phosphate buffer, either 25% of active or inactive CFS was added. The effect on L. monocytogenes was a tenfold reduction. The inhibition of L. ivanovii and L. grayi with a 10 4 -fold and 10 6 -fold reduction respectively was even more pronounced, in phosphate buffer supplemented with the CFS of S. anginosus BSU 1211∆blp3 no reduced survival could be seen (Fig. 6). In contrast, incubation with CFS of S. anginosus BSU 1211 significantly reduced the number of surviving cells (Fig. 6). In a next step, listerial cells treated with 25% CFS of S. anginosus BSU 1211 were analyzed via transmission electron microscopy to investigate the mechanism of action of Angicin. Pictures were thoroughly screened for loss of membrane integrity (Fig. 6). For L. monocytogenes more damaged cells can be seen in the treated group than in the control group. Numbers of intact L. ivanovii cells also seem decreased after treatment with active CFS for 15 min and a lot of cell debris can be seen. Especially for L. grayi membrane disruption is visible in samples treated with active CFS for 15 min. Multiple cells appear lysed and loss of membrane integrity often seems to appear at sites of division. None of these changes can be seen in control cells treated with CFS of S. anginosus BSU 1211∆blp3. Table 2. Bacteriocin stability in cell free supernatant against L. monocytogenes after various treatments. "+" indicates no altered activity, "reduced" indicates reduced activity and "−" means a complete loss of activity, ND indicates not determined. www.nature.com/scientificreports/ Mechanism of action. The most common mode of action of bacteriocins is membrane permeabilization.
To investigate this mechanism in the context of Angicin treatment a SYTOX Green Membrane Permeabilization Assay was conducted. SYTOX Green is a fluorescent DNA stain that can enter the bacterial cell after membrane disruption. Therefore, the fluorescence intensity is an indicator of membrane integrity. Treating L. monocytogenes with 100, 50 or 25 μg/ml synthetic Angicin lead to a significant SYTOX enrichment, indicating membrane disruption of the target cells (Fig. 7). Already after 5 min of incubation with either 100 or 50 μg/ ml synthetic Angicin the membrane is significantly disrupted, indicating a very fast mechanism of action of the peptide (p-value: 0.0079). Additionally, no significant difference between the ethanol treated cells (positive control) and cells treated with 100 μg/ml of Angicin for 15 min can be detected.
Immunity proteins. For self-protection bacteriocin producers express immunity proteins. To explore, if Blp3.3 protein represents as predicted an immunity protein, a functional assay was performed. The gene blp3.3 was introduced into plasmid pAT28 under the control of an endogenous promotor or without a promotor and transformed into the Angicin sensitive target strains S. anginosus SK52 and S. constellatus BSU 1213. In a next step a one-layer RDA was performed and the effect of S. anginosus BUS 1211 on the target strains was determined. Inhibition zones of mutants were compared to inhibition zones of the wildtype strain harboring empty pAT28 plasmid. Transformation of S. anginosus SK52 with blp3.3 under the control of an endogenous promotor led to a significantly reduced sensitivity towards Angicin (Fig. 8). Whereas blp3.3 without promoter did not alter inhibition zone sizes. Comparable results were obtained for the S. constellatus strain BSU 1213. Introducing the blp3.3 gene with an endogenous promotor led to significantly decreased inhibition zones. However, introducing blp3.3 alone into S. constellatus BSU 1213 had no significant effect. In summary, this data supports a role of Blp3.3 in immunity against Angicin. Subsequently, Blp3.3 protein was recombinantly expressed in E. coli and purified to investigate its effect on the antimicrobial activity of Angicin. To that end, we preincubated Blp3.3 together with Angicin for 1 h at 37 °C before the Angicin activity was tested in a two-layer RDA. We found that Blp3.3 alone already caused a zone of inhibited growth (Supplementary Fig. S4a) and it did not decrease Angicin activity on the target strain L. monocytogenes, irrespective of whether we used equimolar, substoichometric or superstoichometric concentrations of Blp3.3. Moreover, recombinant Blp3.3 protein did not affect the inhibitory activity of S. anginosus BSU1211 against L. monocytogenes when analyzing the inhibition zone size in a one-layer RDA. In addition, there was no inhibition of L. monocytogenes by Blp3.3 protein with or without S. anginosus BSU 1211∆blp3 ( Supplementary  Fig. S4b).

Discussion
In their natural habitat, bacteria are part of complex multi-species environments. S. anginosus typically resides in the oral cavity which can be colonized with up to 700 different species 34,35 . Closely related species often need similar nutrients or inhabit a similar space, therefore there is an ongoing competition between these species. Secreting antimicrobial substances, like bacteriocins, provides the producing organism with a great colonization advantage 36,37 .
Putative bacteriocin production of streptococcal species from the S. anginosus group has previously been reported 38,39 . To elucidate the molecular background of this phenomenon, 95 S. anginosus strains were evaluated for inhibitory activity in coculture experiments, which led to the selection of strain BSU 1211 for its potent www.nature.com/scientificreports/ antimicrobial activity. In this strain the genetic basis of bacteriocin production and regulation was further investigated. Bacteriocin production is often regulated through quorum sensing systems 40 . In streptococci, like S. intermedius and S. pyogenes the sil system controls bacteriocin production 12,17 . A sil locus was also identified and predicted to be functional in many of the surveyed S. anginosus strains. In nearly all these strains inducing the sil two component system (SilAB) by addition of the autoinducing peptide SilCR SAG-C , turns on bacteriocin production. In S. anginosus the sil locus is directly adjacent to putative bacteriocin genes, encoded in the blp3 region. By genetic mutagenesis we could show that this region is indeed responsible for bacteriocin production, since a complete deletion mutant of blp3 was unable to inhibit the growth of any target strains. To identify the bacteriocin the three putative bacteriocin genes of this locus were mutated separately. A deletion of blp3.1 and blp3.6 did not alter bacteriocin activity, while deleting blp3.4 completely abolished the inhibitory effect. Interestingly, the deduced peptides of blp3.1 and blp3.6 both contain a GxxxG motif, which is a key feature of two peptide bacteriocins 29 , that function only as a couple. Typically, the structural genes of two peptide bacteriocins are located directly next to each other. But this is not the case in the blp3 region of S. anginosus. It may be possible that due to this spatial gene separation no active bacteriocin can be formed. The putative bacteriocin region of SAG is a known hotspot of genetic diversity 12 . Possibly, the peptides are active when certain environmental conditions are met or other signals are present. The identified S. anginosus bacteriocin was designated Angicin. The Angicin prepeptide is 84 amino acids long of which the first 30 amino acids form the leader peptide. Mature Angicin has a predicted pI of 10.2 and a molecular mass of 6053 Da. In our experiments Angicin showed a good antimicrobial activity against the Gram positive species S. anginosus, S. constellatus, L. monocytogenes, L. grayi, L. ivanovii and VRE. Moderate inhibitory activity could be observed against Gram negative microorganisms such as E. coli, K. pneumoniae, P. aeruginosa, A. baumannii. In some cases, spontaneous Angicin-resistant colonies of S. constellatus, L. monocytogenes and L. ivanovii appeared in the inhibition zones. In further studies, these resistant colonies could be genetically analyzed for mutations and thereby a putative receptor for Angicin might be identified 23 . The appearance of spontaneous bacteriocin-resistant mutants is documented for many bacteriocins 41,42 . This needs to be considered for further applications and to completely inhibit target organisms, often bacteriocins are not administered alone but in combination with other bacteriocins or antibiotics 43,44 .
Comparing peptide sequences, Angicin shows homology to Bovicin (variant 255) and is grouped into the Garvicin Q family which also includes BacSJ as a member 22,[45][46][47] . Garvicin Q is a class IId bacteriocin of Lactococcus garvieae with a wide spectrum of activity. It is active against all tested strains of carnobacteria, enterococci, lactococci, leuconostoc and listeria and most of the tested Lactobacillus and Pediococcus strains. While all species of the genera Bacillus, Campylobacter, Staphylococcus and Streptococcus as well as P. aeruginosa, Salmonella typhimurium and Candida albicans are not susceptible 23 . BacSJ is produced by Lactobacillus paracasei and shows a very similar spectrum of activity with the only exception being a resistance of Lactococcus garvieae against BacSJ but not Garvicin Q 22 . Bovicin (variant 255), a bacteriocin first isolated from Streptococcus gallolyticus, shows only a narrow spectrum of activity with inhibitiory properties against E. faecium, Butyrivibrio fibrisolvens, Lactobacillus ruminis and Peptostreptococcus anaerobius. All Gram negative species as well as Streptococcus bovis, Bifidobacterium thermophilum or Ruminobacter amylophilus are insensitive towards the peptide 47,48 . In summary, Garvicin Q and BacSJ seem to be mainly important for intra-and interspecies competition. This may also be true for Angicin, which inhibits besides other S. anginosus isolates other streptococci, VRE and listeria. www.nature.com/scientificreports/ Bacteriocins mostly kill target cells by membrane permeabilization. The electrostatic interaction between the cationic peptide and the negatively charged bacterial membrane is important as well as amphiphilic structures 49 . However to function more efficiently bacteriocins usually utilize a receptor present in the bacterial membrane 23,50 . While for bovicin variant 255 no receptor is identified yet, it was proven that garvicinQ and BacJS target the IIC and IID components of the mannose phosphotransferase system (Man-PTS) 22,23 . The Man-PTS is described as a receptor for many bacteriocins, mainly the class IIa bacteriocins 24 . Due to sequence homology this gives rise to the question if Angicin might use the Man-PTS as a receptor as well. S. anginosus BSU 1211 is not able to inhibit Gram negative species and synthetic Angicin only shows an inhibition of Gram negative bacteria when administered at high concentrations. As already demonstrated for antibiotics, a reason for bacteriocin ineffectiveness against Gram negative species seems to be the outer membrane 51,52 . It inhibits an interaction between bacteriocin and its cognate receptor. Nisin binding to lipid II for example is inhibited by the outer membrane. However, inhibition of Gram negative species can be enhanced by the addition of chelating agents like ethylenediaminetetraacetic acid, that destroys the outer membrane 53,54 . For Angicin a similar situation may be the case.
Bacteriocins often kill target cells by membrane permeabilization. The electrostatic interaction between the cationic peptide and the negatively charged bacterial membrane is important as well as amphiphilic structures 49 . By a SYTOX Green membrane permeabilization assay it was demonstrated that Angicin leads to a disruption of the membrane in target cells. Already 5 min after Angicin treatment (Fig. 7) a significant difference in fluorescence signal to untreated cells is visible, indicating a fast mechanism of action. A rapid target cell lysis is not unusual and has also been observed in bacteriocins from enterococci and lactobacilli 55,56 . This data is supported by survival assays and TEM analysis, showing reduced cell numbers and rupture of membranes in listeria cells (Fig. 6). However, membrane permeabilization is mainly seen when high concentrations of Angicin are administered. For other bacteriocins it has been speculated that bacteriocins only work as membrane perturbers at concentrations that exceed what is present in a natural environment 57 . In lower or more natural concentrations, bacteriocin producers can use different mechanisms to inhibit competitors. Subtilisin for example interferes with quorum sensing of target organisms and thereby inhibits biofilm formation 58 . Thus, apart from membrane disruption other mechanisms may also play a role in the antibacterial activity of Angicin. For example, Garvicin A has been shown to interfere with cell division by inhibition of septum formation 59 . In CFS treated L. grayi cells loss of membrane integrity often appears at sites of division, which points in a similar direction. For the other listerial species, TEM experiments were not that clear. However, cells look damaged indicating that Angicin may harm target cells not only by membrane disruption but also other mechanisms.
Bacteriocin producers are rendered insensitive towards their own bacteriocin by simultaneous expression of immunity proteins 60 . Immunity proteins have a high specificity towards the bacteriocin and several different modes of action are already described, including loss of bacteriocin binding, sequestering, bacteriocin export or degradation of the bacteriocin [61][62][63][64] . By bioinformatic analysis of the blp3 region, blp3.3 was predicted as an immunity protein. It shows similarities to the immunity protein PisI, which protects against the class IIa bacteriocin piscolin 126 65,66 . Heterologous expression of blp3.3 in Angicin target strains led to a decreased sensitivity of these strains. Similar results were obtained for PisI as well as for other bacteriocins 63,67,68 . Already the introduction of the plasmid alone led to an increased sensitivity towards Angicin. Carrying a plasmid increases metabolic burdens on bacteria, often rendering them more sensitive to stress conditions 69,70 . A possible explanation why introduction of an empty plasmid alone may result in a higher sensitivity towards bacteriocins.
Immunity proteins may form a tertiary complex with the bacteriocin and the bacteriocin receptor 67,71,72 . Formation of this complex seems to depend on conformational changes after the initial binding of the bacteriocin to its receptor. Preincubation of Angicin with blp3.3 did not alter activity, implying that there is no direct interaction between these proteins, which further supports the theory of a tertiary complex. Immunity was increased but not complete upon heterologous expression of blp3.3, which may indicate the presence of a second immunity protein as it is the case for other bacteriocins 73 . Bagle4 predicted blp3.5 as another immunity protein. Furthermore, a CAAX protease is present in the blp3 locus, representing another protein which may be involved in self-immunity against bacteriocins 74 .
Some bacteriocins like nisin, pediocin PA-1 and leucocin A are administered as food preservatives. Angicin inhibits L. monocytogenes, a species posing a major health risk for food safety 75 . For an application in food preservation, characteristics like stability over a wide pH and temperature range are important 76 . With remaining active in a pH ranging from 2 to 10 and at temperatures till 90 °C both criteria are met by Angicin. Furthermore, Angicin shows activity against VRE, a pathogen posing a great health risk 33 . VRE can cause infections like endocarditis, bacteremia and urinary tract infections and is one of the main causes of nosocomial infections 77,78 . Treatment options are limited and new ways to deal with this pathogen are needed 79 . This is the first study describing and characterizing a bacteriocin of S. anginosus. The small cationically charged Angicin not only inhibits closely related species, but furthermore the important food pathogen L. monocytogenes and the clinically relevant vancomycin resistant E. faecium. This activity complemented with high thermal and wide pH stability render Angicin an interesting compound for food preservation and antimicrobial therapies.
General DNA techniques. Commercial

Radial diffusion assay.
To measure bacteriocin activity, a modified radial diffusion assay (RDA) was conducted 80 . For testing antagonism of bacteria against other bacteria a one-layer RDA was performed, whereas for investigating the antimicrobial activity of supernatant or peptides a two-layer RDA was done. In both cases, overnight cultures of target strains as well as putative bacteriocin producer strains (BPS) were centrifuged at 3000×g for 10 min and the pellet was solved in 10 mM phosphate buffer. After repeated centrifugation at 3000×g for 10 min the pellet was reconstituted in 5 ml 10 mM phosphate buffer. Following O.D. 600 nm determination each plate was seeded with 2 × 10 7 bacterial cells of the target strains. For a one-layer RDA bacterial cells were inoculated in still liquid Trypticase Soy Agar (Oxoid) and 20 ml plates were poured. After solidification, holes were cut into the agar with sterile wide bore pipette tips (Axygena corning brand). These wells were filled with overnight cultures of putative BPS, which were adjusted to an O.D. 600 nm of 0.5. Following overnight incubation at 37 °C and 5% CO 2, inhibition zones were measured in cm. In some assays 100 ng of the autoinducing peptide SilCR SAG-C (GWLEDLFSPYFKKYKLGKLGQPDLG) were added simultaneously with the bacteria to test its effect on bacteriocin production. The peptide was synthesized by the Core facility of functional peptidomics (UPEP, Ulm University, Ulm, Germany) based on the deduced sequence of SilCR SAG-C of S. anginosus strain BSU 1211. Purity levels of SilCR SAG-C exceeded 95% and were determined via high performance liquid chromatography (HPLC).
To carry out a two-layer RDA bacterial cultures were inoculated in 15 ml of still liquid 1% agarose and were poured into a petri dish. Wells were punched into the solidified agarose and filled with the test substance. After 3 h of aerobic incubation at 37 °C an overlay with 10 ml Trypticase Soy Agar was performed. Following solidification plates were incubated at 37 °C and 5% CO 2 overnight. Inhibition zones were quantified in cm. This assay was used to assess the antimicrobial activity of differently treated cell free supernatant (CFS). Furthermore, the peptide sequence of Angicin was deduced from the S. anginosus BSU 1211 DNA sequence. In a next step the 54 amino acid sequence without leader peptide (GSGYCKPVMVGANGYA CRY SNGRWDYKVTKGIFQATTD-VIVKGWAEYGPWIPRH) was synthesized by Peptide Specialty Laboratories GmbH (PSL GmbH, Heidelberg, Germany). Angicin was purified using HPLC and afterwards purity was analyzed with HPLC and MS-MALDI. The determined purity was > 98%. The molecular mass was calculated at 6053.1 g/mol, thus 6.05 μg/ml equal 1 μM. It was solved in ultra-pure water. The activity of this peptide was investigated via a two-layer RDA.

Construction of mutants. Construction of blp3.3 mutants.
Based on the plasmid pAT28 two different blp3.3 constructs were created and cloned into E. coli EC101. One vector contained blp3.3 alone and a second vector encoded blp3.3 with its endogenous promoter (promblp3.3). Blp3.3 without promoter was amplified with primers 20/22 and promblp3.3 with primers 21/22. All primers that were used introduced EcoRI and BamHI restriction sites, for cloning purposes. In a next step these constructs were transformed into S. anginosus strain SK52 and S. constellatus BSU 1213 by electroporation as described elsewhere 81 . In brief, bacterial cells were made competent by several washing steps in 10% glycerol. Then 1 μg of plasmid DNA was added and cells were exposed to an electrical pulse. Clones were selected on THY-plates containing 120 μg/ml spectinomycin. Correct insertions were confirmed by nucleotide sequencing.
Construction of blp3 and angicin deletion mutants. A S. anginosus BSU 1211∆blp3 strain was created via splicing by overlap extension PCR (SOE) and competence related transformation as described in Bauer et al. 21 . In brief, blp3 flanking regions of S. anginosus BSU 1211 were amplified with primer pairs 23/24 for Fragment 1 (F1) and 25/26 for Fragment 2 (F2). The primers introduced an overlap to either a lox66 or a lox71 site. The spectinomycin resistance gene was amplified from pGA14-Spc with primers 27/28, which also introduced a lox66 or lox71 site, respectively, adjacent to the spectinomycin gene. Thereby, an overlap of all fragments was achieved. All fragments were fused via SOE-PCR. Subsequently, the linear DNA construct was transformed into S. anginosus strain BSU 1211 by inducing natural competence with CSP-1 21 . BSU 1211 was incubated with 100 ng CSP-1 and after 40 min the DNA construct was added. After two hours of incubation bacteria were plated on www.nature.com/scientificreports/ THY-spectinomycin plates. To eliminate the spectinomycin resistance gene, spectinomycin positive clones were transformed with a cre-recombinase encoding plasmid (pAT18-cre-rec tufA ), which recombines the two lox-sites to a singular lox72 site. By incubating the resulting clones without antibiotics, plasmid loss was induced to create a markerless deletion strain. The same method was used for creating single gene deletion mutants of the blp3 region. For a deletion of blp3.1 primers 5/29 (F1) and 30/31 (F2), for blp3.4 primers 7/32 (F1) and 33/3 (F2), for blp.3.6 primers 34/35 (F1) and 36/37 (F2) were used.
Complementation of S. anginosus BSU 1211∆blp3. To complement the deletion mutant, S. anginosus BSU 1211∆blp3 was transformed with pAT18-blp3 by using the natural competence system as previously described 21 .
The blp3 region was amplified using primers 44/45 and cloned into pAT18. Transformed cells were plated on sheep blood agar plates containing 10 μg/ml erythromycin and incubated anaerobically at 37 °C.

Cell free supernatant (CFS).
Putative bacteriocin producing strains were grown in THY supplemented with 10% FBS for 24 h. After centrifugation for 30 min at 4 °C and either 4000×g (50 ml) or 11,900×g (UZ) (350 ml) a sterile filtration was performed. Cell free supernatant (CFS) was precipitated with 35% Ammonium sulfate under stirring conditions for 1 h at 4 °C. Following centrifugation for 30 min, at 4 °C and 4000×g the supernatant was discarded and the pellet was resuspended in 10 mM phosphate buffer in one tenth of the original culture volume. In a next step, samples were concentrated via 3 K filters (Amicon® Ultra Centrifugal Filters). The concentrate contained the active fraction of the peptide. CFS activity was afterwards tested via a two-layer radial diffusion assay.
Effect of environmental factors on angicin activity. The stability of bacteriocins was investigated in regard to pH, temperature, degradation by enzymes and time. CFS was adjusted to pH values ranging from 2 to 10. Following overnight incubation at the adjusted pH at room temperature (RT), pH was neutralized and the CFS tested for antibacterial activity. Furthermore, CFS was incubated at 40, 60, 80, 90 °C for 1 h or at 90 and 100 °C for 30 min. After cooling down to RT activity was measured. To assess bacteriocin degradation either proteinase K (Sigma) or Lipase of Aspergillus niger (Sigma) was added to CFS to a final concentration of one mg/ml and incubated for 1 h at 37 °C. After inactivation of the enzymes for 1 h at 70 °C and cooling to RT antimicrobial activity was quantified. To further assess stability at low temperatures CFS was stored at 4 °C, − 20 °C and − 70 °C for one year and then inhibitory activity was analyzed. Activity of 50 μg/ml Angicin was tested after incubation at 60, 80 and 90 °C for 1 h. Furthermore, treatment of 5 μg/ml Angicin with 1 mg/ml lipase or proteinase K was conducted and afterwards activity was tested. All measurements were carried out in triplicate. If the test compound was inactive and not used as a control only two independent experiments were conducted. Inhibition assay. In a 96-well plate a serial dilution of CFS from either S. anginosus BSU 1211 or the deletion mutant was prepared in THY broth ranging from 100% CFS to 6.25%. THY broth without CFS supplementation served as control. 5 μl of target bacteria, adjusted to an O.D. 600 nm of 0.1, were added to each well in a final volume of 100 μl. The plate was incubated at 37 °C and 5% CO 2 for 24 h. Absorbance 600 nm was measured every hour for 9 h and after 24 h in a Tecan microplate reader M infinite 200. At least five independent experiments were conducted. Measurements were done in triplicates.

Determination of minimal inhibition
Survival assay. An overnight culture of L. monocytogenes, L. ivanovii or L. grayi was diluted to an O.D. 600 nm of 0.02 in 10 ml THY. When an O.D. 600 nm of 0.1 was reached, 1 ml was transferred to a 1.5 ml Eppendorf tube and centrifuged at 8800×g for 2 min. The supernatant was discarded and the pellet reconstituted in 750 μl 10 mM phosphate buffer with 10% Tryptic Soy broth (TSB). 250 μl CFS of either S. anginosus BSU 1211 or S. anginosus BSU 1211∆blp3 was added and mixed. Cells were incubated at 37 °C and after 0 h, 30 min, 1 h and 2 h serial dilutions were performed and plated on THY. After overnight incubation at 37 °C and 5% CO 2 colony forming units per ml were determined. At least five independent experiments were performed with technical duplicates.
Electron microscopy. Transmission electron microscopy was performed to investigate the effect of active CFS on target bacteria. Therefore, overnight cultures of target bacteria (S. constellatus, L. monocytogenes, L. ivanovii and L. grayi) were adjusted to an O.D. 600 nm of 0.02 and grown to an O.D. 600 nm of 0.1 and then 1 ml was transferred into a 1.5 ml Eppendorf tube and centrifuged at 8800×g for 2 min. The pellet was resuspended in 750 μl 10 mM phosphate buffer with 10% TSB. 25% (250 μl) CFS from either the S. anginosus BSU 1211 or S. anginosus BSU 1211∆blp3 were added and subsequently incubated at 37 °C. S. constellatus and L. monocytogenes were incubated with CFS for 2 h. In contrast, L. grayi and L. ivanovii were only incubated for 15 min, since killing of target cells happend very fast. After centrifugation at 8800×g for 2 min the supernatant was removed. Cells were reconstituted in 40 μl 10 mM phosphate buffer with 10% TSB. Then 40 μl of double concentrated fixing solution was added, consisting of 3.5% glutharaldehyde, 1% Saccharose in phosphate buffer). www.nature.com/scientificreports/ Afterwards samples were postfixed in osmium tetroxide, dehydrated in a graded series of propanol, embedded in Epon, and ultrathin sectioned according to standard procedures. Samples were analyzed with a Jeol 1400 Transmission Electron Microscope. At least 20 images per sample were taken. Experiment was conducted once.
Recombinant expression of the immunity protein Blp3.3. Blp3.3 protein was recombinantly expressed in Escherichia coli RV308. The coding region of Blp3.3 was synthesized (Eurofins) and cloned to the C terminus of a 11 His-tagged maltose-binding protein in a pMAL-C5X vector (New England Biolabs) separated by a cleavage site for tobacco etch virus protease. Protein expression was performed in mineral salt medium M9 by addition of 1 mM IPTG. Protein purification was done in eight steps: (1)

SYTOX green membrane permeabilization assay. A SYTOX Green Membrane Permeabilization
Assay was used to assess the effect of synthesized Angicin against L. monocytogenes. Therefore, an overnight culture was adjusted to an O.D. 600 nm of 0.05 and incubated at 37 °C and 5% CO 2 till an O.D. 600 nm of 0.1 was reached. One ml of the bacterial culture was transferred into an Eppendorf tube and centrifugation at 10,000×g for 2 mins followed. The pellet was solved in one volume of 10 mM phosphate solution with 0.2 μM Sytox (Invitrogen). 90 μl of the bacteria-Sytox solution were mixed with Angicin with final concentrations ranging from 100 to 6.25 μg/ml. A Tecan microplate reader M infinite 200 reader was used to measure fluorescence intensity. Excitation/Emission wavelength were 488/530 nm, respectively. Cells treated with 70% Ethanol for five minutes were used as positive control. Measurements were performed in triplicates and in five independent measurements.
Bioinformatic and statistical analysis. As a source for nucleotide sequences the GenBank database  83 . Analysis of Angicin peptide sequence was performed with Bachem (https:// www. bachem. com/ servi ce-suppo rt/ pepti de-calcu lator/). Endogenous promotor search was carried out with Softberry-BPROM (http:// www. softb erry. com/ berry. phtml? topic= bprom & group= progr ams& subgr oup= gfindb). GraphPad Prism V6 was used to create graphs and do statistical analysis. If not otherwise specified, all experiments were conducted independently at least five times. Experiments with inactive compounds were repeated at least twice. Depicted is always mean ± standard deviation.