We undertook a series of investigations to determine whether lacticin 3147 acts synergistically with a range of clinically important antibiotics. Antibiotics encompassing many families and modes of action were chosen, including cephalosporins, polypeptides, glycopeptides, carbenems, and quinolones. Following this initial screen, it became clear that lacticin 3147 and the polymyxins acted synergistically.
Polymyxins are a group of polypeptide antibiotics that exclusively target Gram negative microorganisms. The five distinct members of this group, polymyxin A-E, were discovered in 1947 and are produced non-ribosomally by different Bacillus polymyxa species . Polymyxin B and polymyxin E (also referred to as colistin), have been used in clinical practice for decades in otic and ophthalmic solutions [12, 13]. Polymyxins are decapeptide antibiotics which consist of a heptapeptide ring, with polymyxin E differing from polymyxin B only by the presence of D-Leu in lieu of a D-Phe. This ring is linked to a tripeptide side-chain which carries an aliphatic chain attached via an amide bond to the amino terminus . The polymyxins carry five positive charges due to the presence of L-α-γ-diaminobutyric acids  and it has been established that the amphiphilic nature of this molecule gives it the ability to interact, bind and traverse the Gram negative outer membrane. The target molecule is lipopolysaccharide (LPS) , and specifically the lipid A component [16, 17]. The polymyxins dissociate protective divalent cations from their association with anionic LPS. This displacement permeabilises the Gram negative outer membrane to allow the polymyxins, or other cationic peptides, to form pores . It should be noted, however, that the use of polymyxins in clinical settings has been restricted to use only where drug resistant pathogens have been encountered. This is due to the toxicity, primarily nephro- and neuro-toxicity, associated with its use , although this toxicity has been suggested to be dose dependent . Nonetheless, the polymyxins are, in many cases, the only antibiotics capable of overcoming specific drug resistant pathogens such as Pseudomonas aeruginosa and Acinetobacter baumannii in cystic fibrosis patients (for reviews see [21–23]). For this reason the polymyxins cannot be ignored, but strategies that could reduce the dose needed for these antibiotics to be effective are highly desirable.
A number of studies have investigated the consequences of combining various antibiotics with polymyxins. Antimicrobial agents such as miconazole , rifampicin [25, 26] meropenem, ampicillin-sulbactam, ciprofloxacin, piperacillin-clavulanic acid, imipenem, amikacin, and gentamicin  ciprofloxacin  trimethoprim, trimethoprim-sulfamethoxazole, and vancomycin , to name but a few, have been the focus of studies to assess if they can work synergistically with polymyxins (also see Yahav et. al., for a review of compounds synergistic with polymyxin E ). To date the only lantibiotic to have been investigated in this way is nisin, which displays synergy with polymyxin B and polymyxin E against Listeria and E. coli[31, 32]. Nisin has also been shown to function synergistically when combined with polymyxin E (and clarithromycin) against Pseudomonas aeruginosa. Combination studies have also recently revealed that lacticin 3147 and the lactoperoxidase system (LPOS) successfully inhibited growth of Cronobacter spp. in rehydrated infant formula . Lacticin 3147, like nisin, is a food grade bactericidal agent obtained from the GRAS organism Lactococcus lactis. Notably, however, it differs from nisin with respect to its target specificity and its greater potency against a number of species . Also the mechanism of action contrasts from the single nisin peptide, in that it requires the interaction of two peptides, Ltnα and Ltnβ, for optimal bactericidal activity.
Here, we report the first study to investigate whether synergy can occur between polymyxin(s) and a two-component lantibiotic. Not only do we reveal that synergy is apparent against a range of strains tested, we also investigated the individual contributions of Ltnα and Ltnβ. We established that, when combined with polymyxin B/E, the levels of lacticin 3147 required to inhibit Gram negative species are equivalent or lower than the levels of lacticin 3147 alone against many Gram positive targets. Thus, in the presence of 0.3125 μg/ml polymyxin B/E, the concentration of lacticin 3147 required to inhibit Cronobacter spp. is less than the lacticin 3147 MIC for Mycobacterium avium subsp. paratuberculosis (MAP) ATCC 19698 or Mycobacterium kansasii CIT11/06 . Similarly the MIC of lacticin 3147 (alone) against many S. aureus (which includes many of the nosocomial pathogens: methicillin-resistant S. aureus (MRSA), S. aureus with intermediate resistance to vancomycin (VISA), S. aureus with heterogenous vancomycin intermediate resistance (hVISA)) [10, 35], is greater than that required to inhibit E. coli species when in the presence of a polymyxin. It is also important to note that synergy with lacticin 3147 may provide a means of reducing the dose of polymyxins required to inhibit specific targets, thereby addressing polymyxin-associated toxicity issues. For example, 8-fold and 16-fold lower levels of the polymyxins are required to inhibit E. coli and Cronobacter when in the presence of lacticin 3147. Furthermore a recent study by Naghmouchi et al., has shown that in addition to its role in providing synergy with polymyxin E, the lantibiotic nisin appears, at certain concentrations, to eliminate its toxicity, as seen in Vero cell lines . Having established the role lacticin 3147 has in polymyxin synergy, further investigations are warranted in order to ascertain if such toxicity preventing attributes are common amongst lantibiotics.
As with previous studies , the solo activities of polymyxin B and polymyxin E against the strains tested here are very similar. With respect to the dual action of lacticin 3147 and polymyxins, it appears that the lacticin 3147-polymyxin B combination has the greater potency against Gram positive targets but that the lacticin 3147-polymyxin E combination has a greater effect against Gram negative strains. Thus, the single amino acid difference between the two polymyxin peptides appears to have an impact on its bactericidal action and target specificity when combined with lacticin 3147. It was also notable that the lacticin 3147 sensitivity of Gram positive microorganisms such as Enterococcus faecium DO, which is already highly sensitive to lacticin 3147, is not enhanced by the presence of the polymyxins. However, in the case of the strains that are relatively more lacticin 3147 resistant, the benefits of adding polymyxin B (especially with respect to Gram positive strains) and polymyxin E (especially for Gram negative strains) is most apparent. It is interesting to note that this phenomenon does not correlate with results obtained during the initial agar based disc assay screen, where the opposite pattern was observed. However, it is acknowledged that the agar-based screen is a much cruder assay, and in that instance polymyxin concentrations were fixed and only lacticin 3147 concentrations were altered. Moreover, no FIC data can be derived and so increased zone sizes may not represent the optimal combination of the antimicrobials as obtained through checkerboard assays. The mechanism by which this synergy occurs with respect to Gram negative targets is presumably based on the action of polymyxin permeabilising the outer membrane to allow lacticin 3147 to gain access to the cytoplasmic membrane and its lipid II target . However, a phenomenon concerning the synergy between polymyxin B/E and the singular peptides Ltnα and Ltnβ is also unveiled during this study. Considering the action of the singular peptides in the absence of polymyxin, a greater quantity of Ltnβ alone, than Ltnα alone, is required to inhibit E. coli (4.7 times versus 1.5 times respectively). This is logical in that Ltnα has been shown to have greater solo activity, and can bind to lipid II and prevent peptidoglycan synthesis . However in the presence of polymyxin B/E, Ltnα needs to be added at a 6 times greater concentration to bring about an inhibitory effect equal to that achieved by Ltnα:Ltnβ combined. In contrast, Ltnβ only needs to be added at a 4.7 fold greater concentration to compensate for the absence of Ltnα and thus Ltnβ seems more potent than Ltnα in the presence of either polymyxin. It is not clear if this is due to the potency of Ltnα being slightly compromised by the activity of the polymyxins or is a reflection of a particularly beneficial interaction between these antibiotics and Ltnβ. Additional studies will be required in order to investigate this further.