*2.4. Isolation of Bacteria from M. producens and Their Antimicrobial Activities* 2.4.1. Isolation and Identification of Bacteria from *M. producens*' Sheath

The collection of *Moorea producens*' biomass at Shimoni (4◦38 55 S, 39◦22 35 E) and Wasini (4◦39 18 S, 39◦22 14 E), Kenya, at low tide was reported previously [19]. Bacterial isolates from the Kenyan *M. producens*' biomass were streaked directly onto marine agar 2216 (10% *w*/*v*). This consistently led to the isolation of colored colonies that were re-streaked to obtain pure strains. The concentration of marine broth was altered between 1% and 10% to differentiate between bacteria growing under poor and rich nutrient conditions, respectively. *Pseudomonas stutzeri* (yellow) was isolated from the sub-culturing of colonies growing together with colonies of *Enterobacter cloacae* (creamy), a mixture that was characterized by a green pigment. There were diverse morphologies of bacterial isolates after overnight incubation including tiny colonies of *Bacillus subtilis*, glassy *Pseudomonas putida*, large soft *Shewanella algae* (purple), and *Pseudomonas stutzeri* (yellow), respectively. Hard colonies were observed, which were identified as *Bacillus licheniformis* (red). A list of representative epibiotic bacterial isolates from the Kenyan *M. producens* is shown in Table 1.

Nearly 70% of all isolates were γ-proteobacteria. Firmicutes were isolated in reasonable quantities (17%), whereas α-proteobacteria and Actinobacteria were minimal. It is possible that the cyanobacterium appropriated pathogenic bacteria of terrestrial origin at low tide. Direct streaking led to the isolation of 137 bacterial strains, from which 24 strains were selected for further study and their organic extracts examined for antimicrobial activities. In common with the *M. producens*' filament, most bacteria isolated from the sheath were found to be close relatives of human pathogenic bacteria.


**Table 1.** A list of representative Kenyan *M. producens* epibiotic bacteria (EB) isolates.

ND–Strain sequences not deposited with Genbank but were inferred from Blast.

2.4.2. Isolation of Brightly Colored Bacteria from *M. producens*' Sheath during Neap Tide

*Moorea producens* was collected from the sampling sites at a time of minimal tide when the sun and the moon were in quadrature. The growing of sheath bacteria onto marine agar by swab cultures showed the Gram-negative *Pseudoalteromonas* sp. to be the dominant genus. Pseudoalteromonas genus belongs to the sulfur-oxidizing symbiont relatives of γ-*Proteobacteria* commonly associated with sea water. However, this dominance is prevailed by treatment of the *M. producens*' sheath with 70% ethanol (*v*/*v*, 2 min) prior to swab culturing, thereby enhancing the growth of brightly colored colonies of bacteria. The 16S rDNA sequencing identified these brightly colored bacteria as *Nocardia cornyebacterioides* (red), *Paracoccus marcusii* (orange), *Micrococcus* sp. (yellow), *Stappia* sp. (pink), and *Bacillus* sp. BacB3, respectively. Additionally, small colonies of unidentified, creamy, Gram-negative rods were isolated. *Nocardia cornyebacterioides* belongs to the genus *Nocardiopsis*. Prior to 16S rDNA identification techniques, the genus *Nocardiopsis* was defined on the basis of chemotaxonomic markers [31] with phospholipids type III and the phospholipids phosphatidylcholine and phosphatidylmethylethanolamine as salient chemotaxonomic features [31,32]. Previously, *Nocardia cornyebacterioides* was reported as an unpublished strain SAFR-015 with accession number AY167850.1 gi:27497670 in the MWG 16S rDNA data-bank. *Paracoccus marcusii* belongs to the α-3 subclass of the *Proteobacteria* [33], whereas the genus *Stappia* belongs to the ecologically important carboxydotrophic bacteria.

Interestingly, the study established that *Pseudoalteromonas* sp. was inhibited by ethanol (70% *v*/*v*) only during the initial 15 min, after which the bacteria population increased linearly with time. This was against the common belief that 70% ethanol has strong sterilizing activity against common bacteria. The natural microbial environment of bacteria on the surface of *M. producens* was mimicked by growing various strains of bacteria in filter discs (6 mm) over a lawn of the dominant *Pseudoalteromonas* sp. Subsequently, these bacterial antagonism experiments showed *Micrococcus* sp. (yellow) and *Stappia* sp. (pink) to antagonize *Pseudoalteromonas* sp. by slowing the growth of the latter bacteria by a zone of 12 mm to 19 mm and of 6 mm to 8 mm, respectively. The diversity of bacteria associated with the cyanobacterium was higher at low tide and lower during neap tide.

2.4.3. Antimicrobial Activities of Organic Extracts of Bacterial Isolates

Using bacteria isolated from the cyanobacteria, organic extracts of Firmicutes were screened against bacteria, fungi, and known pathogens. Organic extracts were comprised of acetone and methanol, whereupon the acetone extracts were clearly the most active and were, therefore, used for the screens. Aliquots of the extracts and a streptomycin control in methanol (1 mg/mL, 20 μL) were tested against Gramnegative *Escherichia coli* strains JW0451-2 and BW25113, respectively, and Gram-positive *Bacillus subtilis* DSM 10 and *Micrococcus luteus* 1790 in 96-well microtiter plates. The absorbances of the organic extracts of microbial isolates were determined by serial dilution in EBS medium (0.5% casein peptone, 0.5% protease peptone, 0.1% meat extract, 0.1% yeast extract, pH 7.0) for bacteria and MYC medium (1.0% glucose, 1.0% phytone peptone, 50 mM HEPES [11.9 g/L], pH 7.0) for yeast and fungi, and were cultivated (30 ◦C, 160 rpm, 24−48 h) with cell concentration adjustment to OD600 0.01 for bacteria and OD548 0.1 for yeast and fungi, respectively. Then, they were matched with absorbances of Streptomycin and Nystatin in IC50 determinations to afford the levels of inhibitions (antibacterial activities), as shown in Table 2. The antibacterial activities were ranked using a percentage scale with the highest activity being assigned a value of 100% (Table 2).

The % scale activity values were obtained as follows:

Absorbance of microbial organic extract Absorbance of control antibiotic at 1 mg/mL×100%

Extracts of *B. marisflavi* JC556 (LS974830.1) were active against *Micrococcus luteus* and *B. subtilis*. Those of *Bacillus licheniformis* PB3 (CP025226.1) displayed moderate activity against *Escherichia coli* and *B. subtilis*, whereas organic extracts of *B. subtilis* MJ01 showed very high activity against *E. coli* and *M. luteus*, respectively. Biological activity was strain dependent, as demonstrated by *B. marisflavi* LQ1 (MG025780.1) that was less active compared to *B. marisflavi* JC556 (LS974830.1). Similarly, *Bacillus subtilis* TBS-CBE-BS01 (MK346244.1) had minor activity against *M. luteus* compared with *B. subtilis* MJ01. Organic extracts of the bacterial isolates showed no biological activities against *Pichia anomala* and *Mucor hiemalis*, and neither were biological activities of the extracts observed against the pathogens *Acinetobacter baumannii*, *Citrobacter freundii*, *S. aureus* Newman, *Pseudomonas aeruginosa* PA14, *Mycobacterium smegmatis,* and the yeast *Candida albicans*. Of the 24 strains, 5 that were considered not interesting due to potential pathogenesis were later dropped and are, therefore, not included in Table 2 above. Organic extracts of *B. marisflavi* JC556 (LS974830.1) and *B. subtilis* MJ01 were the most active. *B. marisflavi* extract was selected for bioassay-guided isolation of its natural products. Streptomycin was the reference antibiotic for bacteria and nystatin was a reference against yeast and fungi, respectively.

*2.5. Isolation of Micrococcin P1 and Micrococcin P2 and Biological Activities* 2.5.1. Isolation of Micrococcin P1 (**1**) and Micrococcin P2 (**2**)

In this study, extracts of a number of cultured Firmicutes were found to inhibit the growth of *Bacillus subtilis* except for those from *B. subtilis*, suggesting that bacteria species can produce antibiotics against pathogens from their own genus. This observation could pave the way for exploring future antibiotics against *Pseudomonas aeruginosa* and *Klebsiella pneumoniae* that are responsible for nosocomial infections from members of the *Pseudomonas* genus and *Klebsiella* spp., respectively. Acetone extracts of *B. marisflavi* cultures were the most biologically active, especially against Gram-positive *Staphylococcus aureus*. The antibacterial compounds identified in these extracts were micrococcin p1 (**1**) and micrococcin p2 (**2**), shown below in Figure 6.


**Table 2.** Antimicrobial activity of bacterial isolates from a marine cyanobacteria.

KEY: 0–24%, No activity; 25–49%, Minor activity; 50–74%, Moderate activity; 75–99%, Active; 100%, Very active. Activities were relative to absorbances of Streptomycin and Nystatin in an IC50 determination.

C-18 HPLC fractionation of the active fraction of *B. marisflavi* using maXis 2 G (BRUKER DALTONICS, Bremen, Germany) afforded the biologically active antibiotics micrococcin P1 (**1**) and micrococcin P2 (**2**) with the retention times of 9.43 min and 9.67 min, respectively (Figure S3). Using HREIMS, the LCMS-MS fragmentation with maXis 4G (BRUKER DALTONICS, Bremen, Germany) unequivocally isolated the known antibiotics micrococcin P1 and micrococcin P2 from *B. marisflavis* (Figures S4–S7). Molecular formulae were identified by including the isotopic pattern in the calculation (SmartFormula algorithm). Both the fractionation with maXis 2G and the high-resolution MS/MS fragmentation using maXis 4G used an acetonitrile/water gradient system that was buffered with ammonium acetate. Details on how the results were acquired can be found in Section 3.7 of the experimental section.

**Figure 6.** Micrococcin P1 (**1**): R1 = OH, R2 = H; Micrococcin P2 (**2**): R1 = R2 = O.

Dereplication of the active molecular ions of 571.60/1142.2 and 572.60/1144.2 using MS2 ions in maXis 4G and the Dictionary of Natural Products linked the strong activity of the *Bacillus marisflavi* strain to its production of micrococcin P1 (**1**) (C48H49N13O9S6; obs. *m*/*z* 1144.20227/572.60381) and micrococcin P2 (**2**) (C48H47N13O9S6; obs. *m*/*z* 1142.20446/571.60370), respectively. These antibiotics were originally isolated from *Micrococcus* sp. [34]. Their structures were confirmed by spectroscopy in 1978 [35] and the total synthesis and stereochemical assignment was accomplished in 2009 [36]. More recent work includes the ribosomal production of micrococcin P1 and micrococcin P2 antibiotics through prepeptide gene replacement [37]. Most recently, the antibiotic thiopeptide micrococcin P3 was isolated from a marine-derived strain of *Bacillus stratosphericus* [38]. In this study, we established the following MS-MS fragmentation pattern for micrococcin P1 (**1**) from the maXis 4G measurements (Table 3) that unequivocally matched with that reported by Walsh et al., in 2010 [37].


**Table 3.** Fragmentation pattern of micrococcin P1 (**1**) 1144.20227/572.60381.

Micrococcin P1 and micrococcin P2 are closely related molecules with the former existing as the secondary alcohol and the latter as the ketone congener. The MS/MS fragmentation pattern shown above for micrococcin P1 is the most common pathway followed by thiocillin variants including micrococcin P2 [37]. It could, therefore, be concluded that together with high-resolution mass spectrometry, dereplication of the active molecular ions, and the Dictionary of Natural Products, bioassay-guided fractionation of the extracts of *B. marisflavi* led to the first isolation of the antibiotics micrococcin P1 (**1**) and micrococcin P2 (**2**) from a strain of *B. marisflavi*. Isolation of these known antibiotics was strain dependent since organic extracts of other strains of *B. marisflavi* isolates did not exhibit the activity of *Bacillus marisflavi* JC556 (LS974830.1).

#### 2.5.2. Biological Activity of Micrococcin P1 against *Staphylococcus aureus*

The activity of micrococcin P1 (**1**) against a panel of pathogens, *Acinetobacter baumannii*, *Citrobacter freundii*, *S. aureus* Newman, *Pseudomonas aeruginosa* PA14, *Mycobacterium smegmatis*, and the yeast *Candida albicans*, was minimal with activity only observed against *Staphylococcus aureus*. This last activity of micrococcin P1 against *S. aureus* was consistent with that in the literature [33,34,39]. MIC is the lowest concentration of the drug, preventing visible growth of the test organism [40], and IC50 is the concentration of the drug at which the growth of half of the test organism (bacteria, fungi or yeast) is inhibited [41]. An MIC value of 0.0157 μM was obtained for micrococcin P1. Using the method of Okanya et al. [41] and its modification as outlined by Jansen et al. [40] and Bader et al. [42], the IC50 of micrococcin P1 against *S. aureus* was shown to be 0.357 μM (0.159 mg/mL) (Figure S8). This value is within the range of IC50 of micrococcin P1 against an assorted collection of *S. aureus* strains and confirms the activity of micrococcin P1 against *S. aureus* to be strain dependent [39]. The IC50 regression curve of the biological activity data of micrococcin P1 against *S. aureus* is shown in Figure S8.
