Mursamacin: a novel class of antibiotics from soil-dwelling roundworms of Central Kenya that inhibits methicillin-resistant Staphylococcus aureus

Bacterial and nematode strains

MRSA strain 133 cultures were obtained as a gift from Dr. John Ndemi of the Kenya Medical Research Institute, Centre for Microbiology Research, Nairobi, Kenya. Pure nematode cultures of Steinernema roundworm isolates were obtained from the nematode culture collection of Horticulture Research Institute Thika, Kenya.

Using the indirect haemolymph method²⁰ X. griffiniae strains XN45 and L67 were isolated from Steinernema sp. Scarpo and Steinernema sp. L67 nematodes respectively. The isolates were cultured on Xenorhabdus differential media NBTA, composed of nutrient agar (Himedia) supplemented with 0.0025% (w/v) bromothymol blue (Sigma-Aldrich) and 0.004% (w/v) 2,3,5 triphenyl tetrazolium chloride (Sigma-Aldrich)²⁰. Identification of the bacteria as Xenorhabdus was based on the presence of the following characteristics: swarming motility on NBTA of 1% (w/v) agar concentration; swimming motility on NBTA of 0.5% (w/v) agar concentration; and yellow green colony pigmentation on NBTA⁵. Specific identification of the bacteria was done by multi locus sequence typing of the 16s rRNA, serC and recA genes (See Dataset 1).

Molecular methods

Total DNA extraction from the bacterial strains was done using FastDNA®SPIN Kit for Soil (MP Biomedicals, USA). Isolation of a 1397 base pair(bp) 16s rRNA gene fragment was done by Polymerase Chain Reaction (PCR) with primer sequences (Inqaba Biotech) as follows: (27f-AGA GTT TGA TCA TGG CTC AG) and (1391r-ACG GGC GGT GTG TGC)²¹. The genes were amplified in a 25 μl reaction volume containing final concentrations of 0.5 U Q5 DNA polymerase (New England Biolabs, USA), 200μM each dNTP, 2mM MgCl_2, and 0.05μM of each primer. Cycling conditions were set at 98°C for 30 s, 40 cycles of 98°C for 30 s, 42°C for 15 s (first 20 cycles) and 47°C for 15 s (final 20 cycles), 72°C for 1 min and a final extension of 72°C for 2 min (MJ Research PTC-100, USA). Isolations of 670bp serC and 400bp recA gene fragments were done by PCR with primers (recA-FW CCA ATG GGC CGT ATT GTT GA) and (recAREV-TCA TAC GGA TCT GGT TGA TGA A) and (serCF-CCA CCA GCA ACT TTG TCC TTT C) and (serCR- AAA GAA GCA GAA AAA TAT TGC AC) respectively²². They were amplified in a 50 μl reaction volume containing final concentrations of 2 U MyTaq^® DNA polymerase (Bioline, USA), 200μM of each dNTP, 3mM MgCl_2, and 0.4μM of each primer. For both, cycling conditions were set at 95°C for 1 min, then 40 cycles of 95°C for 15 s, 52°C for 15 s, 72°C for 40 s, with a final extension of 72°C for 5 min (Thermo Scientific Arktik, USA).

PCR products were visualized on 1.2% (w/v) agarose gels stained with ethidium bromide at final concentrations of 0.5 μg/ml. Typical electrophoresis conditions were 4V/cm for 72 min. Expected bands were excised and purified with Quick Clean II Gel extraction kits® (Genscript, USA). Products were outsourced for sequencing (Macrogen, Netherlands), and obtained sequences were quality checked, assembled and poor quality base calls trimmed in BioEdit²³ and MEGA6²⁴ software suites.

Phylogenetic analysis

Phylogenetic reconstruction was performed using a multi locus concatenate of 16s rRNA, serC and recA genes sequences, that jointly constituted 2076 positions. A dataset of n=13 (1= from this study and 12= public databases) was used that contained the 11 strains of Xenorhabdus and a Photorhabdus luminiscens (the out-group sequence) that had public database sequences of all three genes (see data files). Database sequences were checked for quality and ambiguous nucleotides resolved in the MEGA6 software suite²⁴. Multiple sequence alignments were performed in the same suite using the MUSCLE algorithm²⁵. The evolutionary history was inferred by the ML method based on the Generated Time Reversible (GTR) Model (500 bootstraps). Initial trees for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. The analysis involved 13 nucleotide sequences. All positions containing gaps and missing data were eliminated resulting in a total of 2076 positions used in the reconstruction. Xenorhabdus DNA sequences generated in this study were deposited in the DNA databank of Japan with the partial 16s rRNA gene sequences of X. griffiniae L671, L672, L673, L675, XN45 assigned the following accession numbers respectively; AB987698.1, AB987700.1, AB987701.1, AB987699.1, AB987697.1. The partial recA, serC gene sequences of X. griffiniae L67 and XN45 were assigned LC096094, LC096092 and LC096093, LC096091 respectively. Percentage sequence similarities between gene sequences from this study and other 16s rRNA, recA and serC gene database sequences was determined using blast searches²⁶.

Bacterial fermentation

Fermentation was done using X. griffiniae XN45 bacterial cultures. Multiple colonies (2–3) of an individual isolate were selected, inoculated into 5 ml of Luria Bertani (LB) media containing 1% tryptone, 0.5% yeast extract and 1% NaCl (w/v) and incubated on a rotatory shaker at 150 rpm at 33°C for 24 h. These served as 1% (v/v) starter inocula. Sterile LB media (500 ml) was dispensed into sterile 1-L Erlenmeyer flasks, with starter cultures (5 ml) thereafter inoculated and LB media incubated at 150 rpm at 33°C for 355 h, 180 h and 108.5 h. LB with no inoculum was also incubated to serve as a control for sterility. After fermentation, cells were removed by centrifugation of broths at 20,000 g for 25 min at 4°C (Beckman Avanti J-25, USA) followed by decanting cell free supernatants (cfs). These were heat-treated by autoclaving at 121°C and 15 p.s.i for 20 min to yield a sterile heat stable fraction of the whole broth extract (antibiotic) that was designated “mursamacin”. These were stored at 4°C until use.

Broth macro dilution inhibition assay against MRSA

The broth macro dilution assay was prepared as previously described¹¹ with modifications, using antibiotics obtained from each of the fermentation durations (180.5 h and 355 h) as the antibiotic and MRSA as the test bacterium. MRSA overnight cultures were inoculated into each dilution, at a final concentration of 2.3×10⁴ cfu/ml, and then incubated for 18 h at 37°C without agitation. The following controls were included in every replicate: negative control of 2× LB media inoculated with bacteria without antibiotic; sterility control of 2× LB media with no inoculated bacteria; and sterility control of undiluted antibiotic with no inoculated bacteria.

After incubation, turbidity of each dilution was measured by determining the A600_nm (See dataset files). This was used in the following formula, modified from Houard, Aumelas¹³ that included a correction factor for inhibition by the broth media to calculate the percentage growth inhibition of bacterial cultures by an antimicrobial.

Where g = A600_nm of bacteria in broth culture without antibiotic, and g_x = A600_nm of bacteria in broth culture with antibiotic.

Fractionation of lyophilized powders of mursamacin

To extract an organic heat stable mursamacin class of antibiotics, cfs from a 108.5 h fermentation reaction were lyophilized to yield a yellow powder, whose measured amounts were dissolved in known volumes of methanol, and vortexed for 2 min to yield a dark orange methanol extract. These were centrifuged at 20,000 g for 13 min at room temperature to separate liquid methanol extracts from insoluble residues. Methanol extracts were pipetted into sterile 1.5 ml tubes while pelleted solids were aseptically air dried and then incubated at 37°C overnight to evaporate residual methanol. To provide a negative control, the same procedure was repeated for lyophilized powders of the fermentation media. The difference in weight of powder before and after methanol extraction was then measured to determine the concentration of total dissolved compounds in the methanol extracts.

Broth microdilution assay of organic heat stable mursamacin antibiotics against MRSA

This was modified from a previously described method²⁷. To determine the minimum inhibitory concentrations (MICs) against MRSA, 100 μl of known concentrations of organic heat-stable fractions of mursamacin antibiotics were dispensed into sterile 96-well micro-titre plate starting wells. These were left in a biological safety cabinet to evaporate methanol resulting in visible orange solid residues; each was then re-dissolved in 200 μl RPMI media supplemented with 5% (v/v) LB²⁸. Every other well along each row was filled with 100 μl RPMI media supplemented with 5% (v/v) LB; 100 μl of starting well mixture was then dispensed to the subsequent well resulting halving the antibiotic concentration. This was repeated for all wells along the rows resulting in a 2-fold dilution series. MRSA inocula (10 μl) were used that had been previously prepared from plate cultures dissolved in physiological saline to a turbidity of 0.5 Mcfarland standard (ca. concentration= 2.6×10⁶ cfu/ml), then subjected to 10^-3 dilution in RPMI media supplemented with 5% (v/v) LB. A positive control row of Daptomycin and negative control of methanol extract of LB media only was incorporated in every replicate. Plates were incubated at 37°C for 21 h without agitation. Experiment was performed in 10 replicates in three reproductions.

High performance liquid chromatographic analysis

To identify the compounds contained in the organic heat-stable fraction of mursamacin antibiotics, analytical reverse phase chromatography of the methanol extract was performed as previously described¹³ and modified to use of a c18 column (Aglient Zorbax Eclipse Plus C18; 3.5um, 4.6 ×100 nm) under isocratic conditions of 60% acetonitrile and uv detection of 224 nm.

Xenorhabdus griffiniae strains are found in Kenya

The Xenorhabdus isolates identified were most closely related to Xenorhabdus griffiniae (Figure 1). The apt thresholds for Xenorhabdus species identification based on sequence similarities are currently considered to be 98.65% and 97% for 16s-rRNA and for both recA and serC gene fragments respectively^29,30. In this study, sequences similarities to X. griffiniae strains, including type strains, were 99.52%, 98.57% and 97.68% for 16s rRNA, recA and SerC respectively (See data files)

Figure 1. Phylogenetic tree of Xenorhabdus species based on the maximum likelihood method.

The tree was based on a concatenate of ss-rRNA, recA, and serC gene sequences and was reconstructed with a general time reversal model and test of phylogeny of 500 bootstrap replicates. X. griffiniae isolated from this study (red triangle) clustered in the X. griffiniae clade (red), and was most closely related to X. griffiniae from Malaysia.

Xenorhabdus griffiniae has only previously been isolated from Indonesia³¹, Malaysia³² and South Africa³³; therefore our data strongly suggest a new strain of X. griffiniae originating from Central Kenya.

X. griffiniae produces heat stable anti-staphylococcal compounds

The growth of methicillin-resistant S. aureus was inhibited when cultured in X. griffiniae cell-free supernatants (cfs) that we termed ‘mursamacin’ (Figure 2). Furthermore, cfs were obtained by various X. griffiniae fermentation durations and had the following percentage of growth inhibition against methicillin-resistant S. aureus at neat concentrations: 81% (cfs from 180 h ferment), 94% (cfs from 355 h ferment) (Figure 3). In both instances, cfs were heat-sterilized by autoclaving and pH adjusted to that of the control (6.6–7.0), indicating that inhibition was due to heat stable compounds contained therein.

Figure 2. Inhibition of methicillin-resistant S. aureus by mursamacin antibiotics.

MRSA cultures were incubated with (tube a) and without (tube b) mursamacin antibiotics respectively. A clear tube denotes no bacterial growth while turbid tube denotes bacterial growth.

Figure 3. Growth inhibition of methicillin-resistant S. aureus by heat stable fractions of mursamacin antibiotics produced by varying fermentation durations.

The longer fermentation duration (355 h) produced antibiotics that were generally more inhibitory to methicillin-resistant S. aureus. Dotted line graphs represent linear equations derived from the raw inhibition values. The high R-squared values demonstrate that the concentration of the antibiotic was predominantly responsible (97–98%) for the level of percentage growth inhibition.

Previous studies^{9,11–13,34,35} have demonstrated that Xenorhabdus bacteria are prolific antimicrobial producers with each species producing its unique array of antibiotics that often contained novel compounds. Consistent with these reports, our data demonstrate unprecedented evidence of heat-stable X. griffiniae antimicrobials and further suggests that their production is affected by how long X. griffiniae is cultured in fermentation media.

An organic heat-stable fraction of mursamacin antibiotics is highly effective against methicillin-resistant S. aureus

An organic heat-stable fraction of mursamacin antibiotics inhibited methicillin-resistant Staphylococcus aureus at a standardized concentration of 8.25 μg/ml while its negative control- organic extract of autoclaved fermentation media only- displayed growth at all concentrations confirming this inhibition as due to antibiotic compounds only (Table 1). However, this concentration was 17-fold higher than the positive control Daptomycin that gave 0.5 μg/ml.

Currently, there are two major clinical antibiotics with inhibitory concentrations against methicillin-resistant S. aureus that are considered effective: Daptomycin (0.5 μg/ml) and Vancomycin (2 μg/ml)⁴. On the other hand, the larger majority today’s clinical drugs have inhibitory concentrations against methicillin-resistant S. aureus considered that are considered ineffective; Azithromycin (128μg/ml), Amoxicillin/Clavulanic acid (64 μg/ml), Cefriaxone (64 μg/ml) and Erythromycin (32 μg/ml) and Imipenem (16 μg/ml)⁴. Yet our data demonstrate a heat stable class of compounds with an inhibitory concentration of 8.25 μg/ml. This strongly suggests that antimicrobials contained in this class are even more potent against methicillin-resistant S. aureus, as pure compounds.

Further high performance liquid chromatographic analysis revealed that this fraction contained two major compounds, eluted at 1.862 min and 2.775 min and absorbed at 224_nm when dissolved in methanol (Figure 4). Previous studies have characterized classes of organic Xenorhabdus antibiotics that were highly effective against S. aureus^11,36 and had peaks absorption ranges from 200–230_nm¹⁵. Of note, the PAX lipopeptides¹⁰ isolated X. cabanillasii and X. nematophila were highly effective against MRSA at concentrations of 0.5 μg/ml. Yet in contrast to our results, no class has been characterized as heat stable and isolated from X. griffiniae.

Figure 4. HPLC Chromatographs of organic mursamacin antibiotics.

Top chromatograph represents the solvent only while bottom chromatograph is of solvent containing organic mursamacin antibiotics. Two dominant compounds, indicated by the red arrows, were detected in this fraction.

In conclusion, we demonstrate that X. griffiniae antibiotic compounds termed “mursamacin” contained an organic heat stable fraction are highly effective against methicillin-resistant S. aureus and antimicrobial activity seems to be attributed to two dominant uncharacterized compounds. This may offer a founding stone for further development of clinical drugs from X. griffiniae, giving hope to thousands of patients affected by methicillin-resistant S. aureus infections.