While the bacteriocins characterized from Gram-positive species are predominantly small (< 10 kDa) peptides, several large antimicrobial proteins have been described at both the biochemical and genetic level. These bacteriocins typically manifest as heat-labile proteins, but one apparent exception is propionicin SM1, a heat-stable inhibitory agent produced by P. jensenii (Miescher et al. 2000). It should be noted, however, that aggregates of small peptides, for example, staphylococcin 1580 (Sahl 1994), have caused confusion in the past with regard to estimation of protein size. The bona fide large bacteriocins of Gram-positive bacteria can generally be subdivided into two distinct groups: (1) the bacteriolytic enzymes (or bacteriolysins), which facilitate the killing of sensitive strains by cell lysis, and (2) the non-lytic antimicrobial proteins. However, in some cases, such as propionicin SM1 and albusin B (from Ruminococcus albus; Chen et al. 2004), the lack of mode of action data precludes them at present from placement in this classification scheme (cf. Table 4.3).
4.4.1 Type IIIa: The Bacteriolysins (Bacteriolytic Enzymes) 18.104.22.168 Lysostaphin - The Prototype Bacteriolysin
Originally described more than 40 years ago (Schindler and Schuhardt 1964), lysostaphin (produced by Staphylococcus simulans biovar staphylolyticus)
Table 4.3 The Class III (>10 kDa) bacteriocins of Gram-positive bacteria
Bacteriolytic enzymes (bacteriolysins)
Lysostaphin Staphylococcus simulans subsp. staphylolyticus
Zoocin A Streptococcus equi subsp.
Millericin B Streptococcus milleri
Enterolysin A Enterococcus faecalis
Other large bacteriocins
Helveticin J Lactobacillus helveticus
Prototype and best-characterized bacteriolytic enzyme; plasmid-encoded; comprises two distinct domains separated by a linker; post-translational processing events, mode of action and mechanism of immunity well-defined
First chromosomally encoded (and streptococcal) bacteriolysin to be characterized; attacks the interpeptide crosslink of streptococcal peptidoglycan; similarly to lysostaphin, is composed of two domains; function of zoocin immunity factor (Zif) similar to lysostaphin counterpart (Lif/Epr); biosynthesis of zoocin A may be subject to catabolite repression
Novel broad-spectrum bacteriolysin that cleaves not only the interpeptide crosslinks but also the stem peptide of peptidoglycan
Relatively broad-spectrum bacteriolytic enzyme that may also cleave the stem peptide; C-terminal domain appears very similar to that of bacteriophage lysins
Possibly the first chromosomally encoded and largest non-lytic bacteriocin (37 kDa) to be described; exact mode of action and mechanism of immunity unknown
Plasmid-encoded non-lytic 21-kDa bacteriocin with a narrow spectrum of activity; possesses a novel predicted secondary structure containing a single essential disulfide bond
Schindler and Schuhardt (1964), King et al. (1980), DeHart et al. (1995), Thumm and Gotz (1997), Ehlert et al. (2000)
Simmonds et al. (1996, 1997), Beatson et al. (1998), Lai et al. (2002), O'Rourke et al. (2003)
Beukes et al. (2000), Beukes and Hastings (2001)
Nilsen et al. (2003)
Joerger and Klaenhammer (1986, 1990)
Streptococcus pyogenes M-types 57 and 69
First plasmid-encoded large (17 kDa) streptococcal bacteriocin to be reported; exhibits an unusual inhibitory spectrum, but possesses similar predicted secondary structure to that of dysgalacticin
Simpson and Tagg (1984), Heng et al. (2004)
20-kDa heat-stable bacteriocin identified in the propionic acid bacteria; mode of action and mechanism of immunity unknown
Miescher et al. (2000)
First large bacteriocin to be identified in a ruminai bacterium; mode of action and immunity mechanism unknown; structural gene encodes a prebacteriocin containing an unusually long secretion signal peptide
Chen et al. (2004)
represents the prototype bacteriolysin and is probably the most extensively studied large bacteriocin elaborated by any Gram-positive bacterium. Lysostaphin is a plasmid-encoded glycylglycine endopeptidase that kills sensitive cells by specifically hydrolyzing the pentaglycine crossbridges in peptidoglycan (Robinson et al. 1979; King et al. 1980). A homologue of lysostaphin, ALE-1 from Staphylococcus capitis, has also been characterized (Sugai et al. 1997a). Due to the ability of lysostaphin to kill members of virtually all staphylococcal species, including those that impact on human and animal health, such as St. aureus and Staphylococcus epidermidis, various reports over the years have recommended its use in a variety of medical and agricultural applications (Oldham and Daley 1991; Wu et al. 2003; Shah et al. 2004).
Lysostaphin is initially synthesized as a 493-aa precursor protein (prepro-lysostaphin), which comprises the 246-aa mature form of the bacteriocin plus the following N-terminal extensions: (1) a 36-aa secretion signal peptide at its N-terminus, followed by (2) 195 aa organized into 15 tandem repeats of a 13-aa sequence (Heinrich et al. 1987; Thumm and Gotz 1997). Following export (with concomitant removal of the signal peptide), the tandem repeats are removed by a cysteine protease to yield the fully activated lysostaphin molecule (Neumann et al. 1993; Thumm and Gotz 1997). The lysostaphin molecule is predicted to consist of two distinct domains separated by a short linker sequence: (1) a N-terminal peptidase domain responsible for the catalytic activity of the protein, and (2) a C-terminal targeting domain involved in binding to the peptidoglycan substrate (Wang et al. 1991; Simmonds et al. 1997).
Further studies showed that the genetic determinant conferring immunity to lysostaphin was also located on the plasmid (Heath et al. 1989). The immunity factor, designated lif or epr, encodes a protein that displays homology to the FemAB complex responsible for adding glycine residues to the pen-taglycine crosslinks (Heath et al. 1989; Thumm and Gotz 1997). However, Epr adds serine residues, rather than glycine, and this change in the amino acid composition of the crosslinks is sufficient to protect the cell from the lytic effects of lysostaphin (Robinson et al. 1979; DeHart et al. 1995; Thumm and Gotz 1997; Sugai et al. 1997b; Ehlert et al. 2000). Overall, the findings arising from the studies on lysostaphin and its immunity factor have provided invaluable knowledge not only to researchers working on bacteriocins but also to those trying to elucidate the complexities of cell wall construction.
Within the last 10 years, much progress has also been made in the characterization of bacteriolysins produced by lactic acid bacteria, mainly from members of the genera Streptococcus and Enterococcus. The prototype streptococcal bacteriolytic enzyme is zoocin A, which is specified by a chro-mosomally located gene (zooA) in Streptococcus equi subsp. zooepidemicus. Despite exhibiting limited amino acid sequence similarity, zoocin A and Zif
(the zoocin A immunity factor) share common properties with lysostaphin and Epr, respectively, such as (1) the hydrolysis of streptococcal interpeptide crossbridges, (2) a modular structure consisting of an N-terminal M37-like peptidase domain and a C-terminal substrate-binding domain, and (3) Zif, similarly to Epr, resembles a FemAB-like protein that, when expressed in a heterologous host such as S. gordonii (a zoocin A-susceptible species), confers the expected zoocin-resistant phenotype (Simmonds et al. 1996, 1997; Beatson et al. 1998; Liang et al. 2004). Intriguingly, Zif does not appear to alter the glycine-serine ratios of the interpeptide chain (Beatson et al. 1998), and therefore the exact mechanism of immunity to zoocin A remains enigmatic. A more recent and exciting development is the novel observation that the biosynthesis of zoocin A may be influenced by glucose levels, i.e., it may be catabolite-repressed (O'Rourke et al. 2003). In our laboratory, we have recently identified stellalysin, a new zoocin A-like antimicrobial protein produced by the oral bacterium Streptococcus constellatus subsp. constellatus. Preliminary analyses indicate that stellalysin biosynthesis may also be catabolite-repressed (N.C.K. Heng et al., unpublished data).
Aside from zoocin A and stellalysin, only two other bacteriolytic enzymes produced by lactic acid bacteria have been described, namely, millericin B from Streptococcus milleri and enterolysin A from E. faecalis (Beukes et al. 2000; Nilsen et al. 2003). Millericin B is distinctive in its ability to hydrolyze the cell walls of species such as M. luteus, Staph. aureus and non-millericin B-producing strains of S. milleri, all of which possess different interpeptide crosslinks (Beukes et al. 2000). It was further shown that millericin B could cleave peptidoglycan either in the stem peptide (which is common to the above-listed three species) or in the interpeptide crosslinks (Beukes et al. 2000). Moreover, the mechanism of immunity to millericin B, similarly to that of lysostaphin, involves amino acid substitution (leucine for threonine) in the interpeptide crosslinks of peptidoglycan (Beukes and Hastings 2001).
Enterolysin A is the first large bacteriocin to be described from E. faecalis, and similarly to millericin B, exhibits a rather diverse inhibitory spectrum. Although the common element in the peptidoglycan of all enterolysin A-sensitive species appears to be the stem peptide (L-Ala-D-Glu-L-Lys-D-Ala; Nilsen et al. 2003), the exact mode of action of enterolysin A remains to be determined. Enterolysin A is composed of the two-domain structure typical of other bacteriolysins (Nilsen et al. 2003). Interestingly, while the N-terminal domain of enterolysin A, similarly to that of other bacteriolysins, is of the M37-like peptidase type, the C-terminal domain displays significant homo-logy to the lysins of Lactobacillus casei bacteriophages (Nilsen et al. 2003).
4.4.2 Type IIIb: The Non-Lytic Bacteriocins
As the antithesis to the bacteriolysins, several large bacteriocins have been shown to kill target cells by non-lytic means. This could involve dissipation of the proton motive force, leading to ATP starvation and ultimately cell death.
The first non-lytic bacteriocin to be described at the biochemical and genetic level was helveticin J, a 37-kDa bacteriocin produced by Lactobacillus helveticus that primarily targets other Lactobacillus species (Joerger and Klaenhammer 1986, 1990). However, the precise mode of action of helveticin J remains unknown.
Dysgalacticin (21 kDa) and streptococcin A-M57 (SA-M57; 17 kDa) are secreted bacteriocins produced by Streptococcus dysgalactiae subsp. equi-similis and M-type 57 S. pyogenes, respectively (Wong et al. 1981; Heng et al. 2004, 2006). The inhibitory spectrum of dysgalacticin is fairly narrow and is limited to strains of Lancefield serogroups A, C and G (Wong et al. 1981; Tagg and Wong 1983). On the other hand, the range of organisms inhibited by SA-M57 is unusual, consisting mainly of non-streptococcal Gram-positive species including M. luteus, L. lactis, all tested species of Listeria (including Lis. monocytogenes), Bacillus megaterium and St. simulans (Simpson and Tagg 1983; Heng et al. 2004). Both bacteriocins appear to kill sensitive cells in a non-lytic fashion (Wong et al. 1981; Simpson and Tagg 1983; Heng et al. 2006), although the exact mechanism remains unclear.
At first glance, the similarities between dysgalacticin and SA-M57 appear superficial (Heng et al. 2004, 2006): (1) the structural genes for both dysgalacticin (dysA) and SA-M57 (scnM57) are plasmid-borne, and (2) both bacteriocins are exported via Sec-dependent systems. Dysgalacticin does not display any similarity either to proteins of known function or to hypothetical proteins in publicly available databases. Conversely, SA-M57 exhibits primary amino acid sequence similarity with two hypothetical, potentially secreted proteins, EF1097 and YpkK, from E. faecalis and Corynebacterium jeikeium, respectively (Heng et al. 2004).
Despite the obvious lack of sequence similarity between dysgalacticin, SA-M57, EF1097 and YpkK, all four proteins possess similar predicted secondary structures consisting of (1) a fairly unstructured N-terminal portion,
(2) a C-terminal region that appears to contain a helix-loop-helix motif, and
(3) two cysteine residues that are predicted to form a disulfide bond. We have subsequently shown, for both dysgalacticin and SA-M57, that the two cys-teines do indeed form a disulfide bond essential for antimicrobial activity (N.C.K. Heng et al., unpublished data). Furthermore, we have successfully expressed the EF1097 and YpkK structural genes in E. coli, and found that both recombinant proteins exhibit antimicrobial activity, with the former displaying a much broader inhibitory spectrum (P.M. Swe and H.J. Baird, unpublished data). Taken collectively, dysgalacticin, SA-M57, EF1097 and YpkK potentially constitute a novel family of antimicrobial proteins.
Was this article helpful?