In Vitro Activity against Staphylococcus aureus of a Novel Antimicrobial Agent, PRF-119, a Recombinant Chimeric Bacteriophage Endolysin^,

Abstract

Antistaphylococcal activity of the novel chimeric endolysin PRF-119 was evaluated with the microdilution method. The MIC₅₀ and MIC₉₀ of 398 methicillin-susceptible Staphylococcus aureus isolates were 0.098 μg/ml and 0.391 μg/ml, respectively (range, 0.024 to 0.780 μg/ml). Both the MIC₅₀ and MIC₉₀ values of 776 methicillin-resistant S. aureus isolates were 0.391 μg/ml (range, 0.024 to 1.563 μg/ml). All 192 clinical isolates of coagulase-negative staphylococci exhibited MIC values of >50 μg/ml. In conclusion, PRF-119 exhibited very good activity specifically against S. aureus.

TEXT

Prevention and treatment of infections caused by Staphylococcus aureus have become more challenging with the worldwide emergence of methicillin-resistant S. aureus (MRSA) strains. The human anterior nares are the principle habitat for S. aureus as part of a complex microbiota, and nasal carriage has been shown as the origin and a major risk factor for subsequent infection (28, 32, 35). Consequently, nasal decolonization approaches have a major impact on MRSA prevention strategies (1). Although nasal administration of mupirocin ointment results in a statistically significant reduction in S. aureus infections (27), new antibiotics are needed for nasal decolonization because increasing resistance to mupirocin has to be considered (5, 20). Recently, the dramatic increase in antibiotic resistance revitalized interest in use of recombinant phage endolysins and enzymatic bacteriocins to combat pathogenic bacteria (8). Phage-derived endolysins are bacterial cell wall (peptidoglycan)-degrading enzymes that are produced by bacteriophages at the final stage of multiplication to release their progeny (16, 36). They exhibit rapid and strong bacteriolytic activity even when applied exogenously (15). High efficacy and safety of recombinant lysins against S. aureus were demonstrated in vivo in mice by Rashel et al., who showed efficient nasal elimination of MRSA (22). Moreover, intraperitoneal administration of lysins protected mice from septic death after the inoculation of MRSA into the peritoneal cavity (22). Thus, the rapid activity and high specificity of lysins make them attractive target-specific antibacterial agents against antibiotic-resistant pathogens, including MRSA.

PRF-119 (owned by Hyglos GmbH, Regensburg, Germany) represents a recombinant chimeric lysin containing two distinct functional modules: an enzymatic active domain (EAD), namely, the cysteine- and histidine-dependent aminopeptidase/hydrolase (CHAP) domain from the endolysin of phage K, and a cell wall binding domain (CBD) from the bacteriocin lysostaphin. Phage K is a polyvalent phage of the Myoviridae family that is active against a wide range of staphylococci (19). Lysostaphin secreted by Staphylococcus simulans is a bacteriolytic enzyme that cleaves the pentaglycine cross-bridges of staphylococcal peptidoglycans, especially that of the target organism, S. aureus (14, 24), while the peptidoglycans of coagulase-negative staphylococci (CoNS) are usually less susceptible to lysostaphin (37). Cell wall-degrading bacteriocins are released by bacteria to inhibit the growth of similar or closely related bacterial strains (14).

Here, the activities of PRF-119 against methicillin-susceptible S. aureus (MSSA) and MRSA were determined, and its specificity was investigated by testing a large collection of well-defined clinical isolates and reference strains of non-S. aureus staphylococcal species, including both CoNS and coagulase-positive staphylococci other than S. aureus.

(This work was presented in part at the 50th Interscience Conference on Antimicrobial Agents and Chemotherapy, Boston, MA, 12 to 15 September 2010 [F1-2080].)

A total of 1,169 S. aureus isolates comprising 395 MSSA and 774 MRSA isolates were tested. MSSA isolates were obtained from patients during the course of one monocenter or one multicenter study that included 32 university and community hospitals (28). MRSA isolates were collected during the course of a German multicenter study (36 centers comprising university and general hospitals and outpatient clinics) covering 132 different spa types (30). Only one isolate per patient was included. Species affiliation was confirmed by detection of the S. aureus-specific nuc gene as described elsewhere, and methicillin resistance was determined by mecA gene detection (3, 18). In addition, 192 isolates of CoNS from 21 centers in Germany (29) were tested, and they encompassed 165 Staphylococcus epidermidis isolates, 19 Staphylococcus haemolyticus isolates, and 8 other CoNS isolates belonging to Staphylococcus capitis subsp. capitis (n = 3), Staphylococcus warneri (n = 2), Staphylococcus lugdunensis (n = 1), Staphylococcus saprophyticus subsp. saprophyticus (n = 1), and Staphylococcus sciuri (n = 1). Also, 80 staphylococcal reference strains were included comprising 5 S. aureus strains (Table 1) and 75 non-S. aureus strains (see Table S1 in the supplemental material). Four bacteriophage-resistant S. aureus strains, which were recently isolated from the phage-sensitive parental strains for the development of a vaccine active against S. aureus (4), were also tested in this study, as were their respective ancestral strains. To further characterize PRF-119, the activities against SA113ΔtagO, an S. aureus mutant lacking polyribitol phosphate (RboP) wall teichoic acid (WTA) (31), as well as the parental wild-type strain, SA113, were investigated. Furthermore, a polyglycerol phosphate (GroP) WTA-expressing S. aureus strain (PS 187) (7), which represents a distant S. aureus lineage often isolated from dogs (21), was tested.

Table 1.

Antimicrobial activity of PRF-119 against S. aureus

Organisma (no. of isolates tested)	MIC data (μg/ml)
	MIC50	MIC90	Range
MSSA (398b)	0.098	0.391	0.024–0.780
MRSA (776c)	0.391	0.391	0.024–1.563

^aThe exact colony number of the inoculum (mean ± SD) as determined by a surface culture method constituted on average 3.31 × 10⁵ CFU/ml ± 1.62 × 10⁵ with a range of 0.71 × 10⁵ to 7.88 × 10⁵ for MSSA isolates and, for MRSA isolates, an average 3.22 × 10⁵ CFU/ml ± 1.64 × 10⁵ with a range of 0.7 × 10⁵ to 7.98 × 10⁵ (1 × 10⁵ to 5 × 10⁵ CFU/ml is recommended by the German DIN guideline).

^bIncluding three MSSA reference strains: ATCC 29213 (MIC, 0.391 μg/ml), ATCC 25923 (MIC, 0.780 μg/ml), and ATCC 19095 (MIC, 0.098 μg/ml).

^cIncluding two MRSA reference strains: ATCC 43300 (MIC, 0.391 μg/ml) and ATCC 33592 (MIC, 0.585 μg/ml).

The MICs were determined by means of the broth microdilution method according to the German Deutsches Institut für Normung (DIN) 58940-8 guideline (6). A stock solution of PRF-119 was obtained from Hyglos GmbH and stored frozen at −20°C in a storage buffer (25 mM HEPES, 300 mM arginine, 150 mM NaCl, 10 mM CaCl₂; pH 8) and was thawed at room temperature around 20°C before dilution to prepare final test concentrations from 0.000095 to 50 μg/ml. Larger 4-fold dilution intervals were chosen for all dilution steps except the first one (2-fold), in order to cover possible extreme MIC values. For isolates that demonstrated the highest MIC values among all isolates tested, additional tests with smaller concentration intervals were performed to better determine the upper limit of the MIC range. Sterile 96-well microtiter plates (Greiner Bio-One GmbH, Germany) were used. The turbidity of the overnight broth culture was adjusted by using a photometer. The microdilution plates were inoculated to achieve approximately 1 × 10⁵ to 5 × 10⁵ CFU/ml. The exact colony number of the inoculum was determined by a surface culture method. For this, inoculum dilutions were plated onto the agar medium and colonies were counted. The reading was done after 24 h of incubation by using a photometer for microtiter plates at 595 nm and confirmed visually. In deviation from the German DIN guideline, brain heart infusion broth was used and an incubation temperature of 30°C was chosen, reflecting the temperature on the surface of the anterior nares (33). Since these two conditions deviated from the guideline, we performed additional experiments comparing the impacts of different incubation temperatures and broths on PRF-119 susceptibility testing. It was shown in a representative number of strains, including reference strains, that the different incubation temperatures (30°C versus 36°C) and different broths (brain heart infusion broth versus cation-adjusted Mueller-Hinton broth) did not change the MIC results considerably (see Tables S2 and S3 in the supplemental material). All tests were performed at least in duplicate. Additionally, sterility and growth controls were performed.

The MIC distribution data for MSSA and MRSA isolates are shown in Table 1. PRF-119 exhibited excellent activity against S. aureus, with no marked difference between MSSA and MRSA isolates.

While the CHAP domain of phage K should in principle be active against most staphylococcal species, the inactivation of S. aureus by PRF-119 probably relies on the specificity of the CBD from lysostaphin. Binding studies with the CBD of endolysin ALE-1 have shown that a stretch of 9 amino acids at the N terminus of the CBD confers specificity to the S. aureus cell wall (17). This specificity-determining stretch is conserved also in the CBD of lysostaphin (9) and is part of PRF-119.

Point-like growth in wells with agent concentrations above the MIC was observed for some isolates, and this resembled the “trailing effect” known from susceptibility testing of staphylococci with linezolid (12, 26) and fosfomycin (10) and of Candida spp. testing with azole antifungal agents (2). However, this effect was inconstant and often not reproducible. Only in one MSSA isolate and four MRSA isolates was this effect stable in repeated tests. We categorized those isolates as susceptible because there was a distinct difference between turbid growth in wells with agent concentrations below the MIC and minimally visible point-like growth in wells with agent concentrations above the MIC. The mechanism of this “minitrailing” phenomenon is unclear and warrants further examination. Probably, this effect is caused by enhanced intercellular adhesion. A macroscopically visible aggregation of S. aureus cells after adding an endolysin was reported earlier by Sass et al. (23). A putative generation of protoplasts was excluded by the detection of the bacterial cell wall from “trailing” colonies by Gram staining (data not shown). The relevance of the trailing effect in testing staphylococci has been categorized as questionable (10, 12) and will have to be investigated for PRF-119 in further studies.

CoNS isolates of a tested clinical collection were unsusceptible to PRF-119. Even at the maximum concentration tested, growth was observed for all CoNS isolates included (MICs, >50 μg/ml). The exact colony number of CoNS isolates amounted on average to 3.07 × 10⁵ CFU/ml ± 1.90 × 10⁵ (average ± standard deviation [SD]; range, 0.08 × 10⁵ to 9.34 × 10⁵) and was in general more variable than for S. aureus isolates, due to the growth peculiarities of different species. The species presented in this clinical collection cover the most common CoNS species colonizing the human body, thus indicating that the effect of PRF-119 if administered to patients would be quite specific. For some CoNS isolates, a change from diffuse turbidity to growth in the form of plaques was observed at higher PRF-119 concentrations.

Testing of staphylococcal type and reference strains confirmed the activity of PRF-119 against S. aureus, while most of the CoNS species showed no susceptibility, as demonstrated for clinical CoNS isolates. PRF-119 showed activity against some CoNS type and reference strains that were not part of the clinical isolate selection, e.g., S. auricularis, S. chromogenes, S. hominis subsp. hominis, and S. schleiferi subsp. schleiferi (see Table S1 in the supplemental material). For some other species, e.g., S. lugdunensis, S. capitis, and S. saprophyticus, intraspecies variation was observed. Killing of staphylococcal species other than S. aureus by lysins derived from S. aureus bacteriophages as well as by lysostaphin has already been reported elsewhere (22, 24). It is noteworthy that PRF-119 demonstrated activity against type and reference strains of all other coagulase-positive/-variable staphylococcal species known to be associated with different animal hosts, such as S. delphini, S. hyicus, S. intermedius, S. lutrae, S. pseudintermedius, and S. schleiferi subsp. coagulans.

The rate of developing resistance to phages is much slower than that to conventional antibiotics (25). S. aureus mutant strains, which are resistant to certain temperate phages, were recently generated in vitro for the development of an S. aureus vaccine (4). PRF-119 demonstrated activity against these four mutants (A172, A178, A180, and A182) (4), showing MICs of 0.391 μg/ml, 0.098 μg/ml, 0.098 μg/ml, and 0.391 μg/ml, respectively. Their respective ancestral strains showed the same MICs, excluding the ancestral strain of A178, which exhibited only one dilution step difference to the resistant strain. The molecular basis of phage resistance in the four mutants has remained unclear. Our observation that PRF-119 is active against these strains indicates that the CBD may use another ligand than the receptor binding proteins of the phages used to identify the mutants, thereby supporting exchange strategies for CBDs or EADs (“domain swapping”) to generate new recombinant lysins to avoid phage resistance.

PRF-119 demonstrated activity against wild-type S. aureus strains SA113 (MIC, 0.098 μg/ml) and PS 187 (MIC, 0.391 μg/ml), which express RboP and GroP WTA, respectively. PRF-119 was also active against the WTA-lacking mutant SA113ΔtagO (MIC, 0.391 μg/ml); however, a “trailing”-like effect as described above for some other isolates was reproducibly observed with this mutant. Subcultured SA113ΔtagO cells from those microtiter plate wells exhibiting the “trailing”-like phenomenon showed again the same “trailing”-like growth pattern as the initial culture. In a previous study, the S. aureus phages 3A, 52, and φ11, which employ WTA as a receptor, were shown to be completely inactive against ΔtagO (31). Although WTA has been shown not to be the ligand of the lysostaphin CBD (11), its absence seems to have some impact on the activity of PRF-119. We can currently only speculate why PRF-119 is specific for S. aureus and largely inactive against most CoNS isolates. S. aureus is distinguished by production of RboP WTA, while most CoNS isolates produce GroP or more complex types of WTA (7). It is possible that only RboP WTA allows efficient access of the CBD to peptidoglycan, while other types of WTA interfere with its binding. This notion is supported by the fact that the unusual GroP WTA-producing S. aureus PS 187 was less susceptible to PRF 119 than many other S. aureus strains. In addition, many CoNS strains contain serine or alanine residues in the peptidoglycan interpeptide bridges, which reduces the susceptibility to lysostaphin, while the S. aureus interpeptide bridges are composed of five glycines (34). For phage endolysins, it has been reported that an amino acid residue change may contribute to enhanced cell wall-hydrolyzing activity and lower MICs against S. aureus strains (13).

In summary, recombinant endolysin PRF-119 was highly active against all S. aureus isolates tested, including MRSA isolates, while CoNS isolates from clinical collections were not susceptible. By avoiding development of resistance and adverse effects on the normal indigenous flora, as observed for traditional antimicrobials, the use of this chimeric lysin may offer an alternative strategy for MRSA decolonization.