*2.3. Bacterial Isolates from M. producens' Filaments*

In the quest to establish if *M. producens*' filaments could have their associated bacteria completely removed, detached filaments were stained with acridine orange to establish cyanobacteria-bacteria association of the filament and with nigrosin to study cell wall and DNA degradation of the filament by bacteria.

Clearly, despite treatment of the cyanobacteria with cycloheximide, several rinses with phosphate buffered saline (PBS) and nitrogen liquefaction, microscopy and the isolation of the live filament bacteria (LFB) and dead filament bacteria (DFB) from the filament surface indicated that bacteria are always associated with *M. producens* (Figure 3A). Treatment of the cyanobacteria biomass with CuS04·5H2O killed all bacteria except *Pseudoalteromonas carrageenovora* and *Ochrobactrum anthropi*. These surviving bacteria associated with the filament were plated on marine agar and identified through 16S rDNA gene partial sequence. The association of bacteria with cyanobacteria was regardless of the cyanobacteria getting actively involved with phototropism or on the verge of dying. Oxygen is acknowledged to be poisonous towards cyanobacteria including *M. producens*. Bacteria utilize oxygen for respiration and cyanobacteria capture carbon dioxide during photosynthesis, creating an energy balance on the cyanobacteria–bacteria interface. This close metabolic dependency may explain the close physical association of bacteria with the protective sheath surrounding the *M. producens*' filaments. When bacteria on the surface of an untreated filament were left overnight to die, it was observed that they did not enter the core of the filament targeting the DNA material. Nevertheless, there was considerable loss of cell wall material (Figure 3B), leading to the speculation that some of the bacteria on the surface survive on organic carbon from the cyanobacteria. These findings overnight were corroborated by a broken cell wall, as observed on a nigrosine-stained filament similarly left to die.

**Figure 3.** Bacteria on the surface of a live *M. producens*' filament (**A**) and DNA in blue at the center and cell wall material outside (**B**). The images of the filaments on a # 1.5 (~170-μm-thick) glass cover slip were acquired by the author using a Leica DMIRB inverted microscope fitted with a color and black-and-white cooled digital camera. Acridine orange was used to stain the filament shown in (**A**) and nigrosin stained the filament shown in (**B**). Scale bar = 10.36 μm.

Most bacteria isolated from the filaments of *M. producens* are close relatives of human pathogenic bacteria. The presence of Firmicutes and Actinobacteria on the filaments that are ubiquitous with antimicrobial agents suggested that within the consortium some bacterial species, in addition to other roles, may provide chemical defense molecules, which benefit the bacterial cells as well as the host cyanobacterial cells. However, it was unclear how the living filament retained its cylindrical shape, despite the presence of many species of bacteria, some of which were cellulose degraders, as exhibited by the bacteria on a dying filament (Figure 4B).

**Figure 4.** A non-uniform near-dead filament with congregation of live bacteria on the dead filament shown by the red color (**A**). Bacteria embedded in surface of a dying filament (**B**). The green and brown areas in (**A**,**B**) are the live and dead parts of the filament, respectively. The images were acquired by the author, as in Figure 3 above. Staining of the filament was with acridine orange. Scale bar = 10.34 μm.

Further, microscopy of the Kenyan *M. producens* also showed that this cyanobacterium is associated with stalked diatoms (Figure 5). The apparent presence of stalked diatoms on the *M. producens*' filaments indicated that microbes from estuarine and freshwater habitats may have colonized their surfaces [30]. Stalked diatoms, in particular, are found in marshes, rivers, and lagoons and are characteristically in freshwater. In the present study, *M. producens*' samples were taken from an area around the Shimoni creek. This creek is a freshwater river inlet into Shimoni Bay.

**Figure 5.** Filament of the Kenyan *M. producens* with stalked diatoms protruding from both its sides. Light microscopy of the filament with an OLYMPUS BX 51 equipped with inter-differential contrast optics and digital camera. The image was acquired by the author. Scale bar = 20.53 μm. Stalked diatoms were confirmed as being of a terrestrial nature by Prof. Brian Whitton of Durham University, United Kingdom.
