UV Resistance of bacteria from the Kenyan Marine cyanobacterium Moorea producens

Abstract UV resistance of bacteria isolated from the marine cyanobacterium Moorea producens has not been observed previously, findings which highlight how unsafe germicidal UV irradiation for sterilization of air, food, and water could be. Further, UV resistance of Bacillus licheniformis is being observed for the first time. This study focused on bacteria isolated from the marine cyanobacterium M. producens collected off the Kenyan coast at Shimoni, Wasini, Kilifi, and Mida. UV irradiance of isolates (302 nm, 70 W/m2, 0–1 hr) established B. licheniformis as the most UV resistant strain, with the following order of taxon resistance: Bacilli> γ proteobacteria > Actinobacteria. UV resistance was independent of pigmentation. The maximum likelihood phylogenetic distance determined for both B. licheniformis and Bacillus aerius relative to M. producens CCAP 1446/4 was 2.0. Survival of B. licheniformis upon UV irradiance followed first‐order kinetics (k = 0.035/min, R 2 = 0.88). Addition of aqueous extracts (2, 10, 20 and 40 mg/ml) of this B. licheniformis strain on the less resistant Marinobacterium stanieri was not significant, however, the commercial sunscreen benzophenone‐3 (BP‐3) positive control and the time of irradiance were significant. Detection of bacteria on M. producens filaments stained with acridine orange confirmed its nonaxenic nature. Although the chemistry of UV resistance in cyanobacteria has been studied in depth revealing for example the role of mycosporine like amino acids (MAAs) in UV resistance less is known about how bacteria resist UV irradiation. This is of interest since cyanobacteria live in association with bacteria.

Cyanobacteria fulfill vital ecological functions in the world's oceans, being important contributors to global carbon and nitrogen budgets. They are arguably the most successful group of microorganisms on earth, having existed for the last 3 billion years.
They are genetically diverse; occupy a variety of niches including habitats across all latitudes, marine, and terrestrial ecosystems and are found in extreme environments such as hot springs, salt works, and hypersaline bays. In response to radiation, cyanobacteria utilize photoprotective compounds such as scytonemin and its derivatives, mycosporine-like amino acids (MAAs) with high molar extinction coefficients to resist UV radiation (Gao & Garcia-Pichel, 2011;Siezen, 2011;Torres et al., 2004). Mycosporines and MAAs are UV-absorbing small molecules (λ max = 310-360 nm) and are also synthesized by fungi and eukaryotic micro-and macroalgae (Bandaranayake, Bemis, & Bourne, 1996;Shick & Dunlap, 2002).
Their success in preventing UV-induced skin damage in vivo has led to their commercialization, for example in the products Helioguard 365 and Helionori sunscreens to protect against UV-A (Balskus & Walsh, 2010;De la Coba et al., 2009;Siezen, 2011).
A sustainable supply of such sunscreen molecules and other bioactive cyanobacterial compounds could be achieved through aquaculture, chemical synthesis or by recombinant biosynthesis from a source organism. However, obtaining sunscreens through aquaculture is not currently economically feasible as cyanobacteria have a complex circadian rhythm compared with bacteria and it is expensive to grow them (Kondo & Ishiura, 2000). Low enantiomeric excess (e.e) yields and refractory problems associated with chemical synthesis are further limitations to realising sustainability.
Fossil evidence predates 440 Ma for mutually beneficial functional and metabolically interactive associations between cyanobacteria and heterotrophic bacteria (Tomescu, Honegger, & Rothwell, 2008). Bacteria routinely associate with cyanobacteria for buoyancy, organic carbon, nitrogen, and sheath nutrients. In return cyanobacteria benefit from increased photosynthesis arising from removal of oxygen sequestered by bacteria during symbiotic growth (Paerl, Pinckney, & Steppe, 2000). However, much research on bacteria-cyanobacteria interactions has been conducted, virtually none has focused on whether bacteria can contribute to the UV resistance of their host cyanobacterium (Hube, Heyduck-Söller, & Fischer, 2009;Salomon, Janson, & Granéli, 2003;Tuomainen, Hietanen, Kuparinen, Martikainen, & Servomaa, 2006). It is also possible that cyanobacteria provide UV protection to symbiotic strains as well as symbionts enhancing the UV protection of cyanobacteria. Bacteria are easy to grow and could offer sustainable yields of photoprotective agents for commercial application.

| Collection and identification of M. producens
Bacteria in this study were isolated from samples of the filamentous marine cyanobacteria M. producens identified previously (Davies-Coleman et al., 2003). Moorea producens specimens were collected from Kilifi (039.785˚E to 039.835˚E) and Mida Creek (039.99505º to 039.96600ºE) on the North Coast; and at Shimoni (039.36565ºE to 039.36696ºE) and Wasini (039.35906ºE to 039.35942ºE) on the South Coast of Kenya in 2011. The choice of location was based on the ubiquitous availability of the cyanobacteria and on geographical positioning. M. producens specimens were handpicked from a 100 m transect along the shore line to a water level of nearly 0.5 m from the shore.
M. producens samples were thereafter placed in sterile polythene bags and transported to the laboratory at Pwani University, Kilifi for storage at 4°C prior to workup. A second collection of M. producens at Shimoni was also made in 2012 and the specimens were similarly treated before transportation to the United Kingdom for analysis.
During workup in the laboratory, M. producens was treated with cycloheximide (5 mg/L, overnight) to reduce contaminating eukaryotic cells, protozoa, and fungi. It was then rinsed several times with filtered sterile seawater (12 times, 45 μl) and left overnight in phosphate poor autoclaved seawater. For detachment of filaments, the cyanobacterium was submerged in phosphate buffered saline (PBS, pH 7.4), filaments were rinsed with sterile water, pooled together and aliquots weighed to provide sufficient biomass for microscopy, DNA extraction, and genome sequencing. Filaments were stained with acridine orange and observed, using a Leica DMRB microscope with a Micropublisher digital camera (Figure 1). F I G U R E 1 Acridine orange stained bacterial cells on the surface of a live Moorea producens filament

| Genomic isolation and identification of bacterial isolates
Bacterial isolates were streaked onto marine agar 2216 (Difco) for strain isolation. Pure isolates were obtained after successive streaking. Bacterial isolates were cultured in marine broth 2216 (50 ml, Difco) to generate biomass for DNA extraction. Cultures were incubated (28°C) with continuous shaking for 3-5 days after which they were harvested by centrifugation. DNA was extracted, using standard phenol-chloroform procedures (Sambrook & Russell, 2001).
The PureLink R genomic DNA kit from Invitrogen was used according to manufacturer's instructions. Specifically, bacterial cells (200 mg) were treated with lysis buffer (1 ml) and incubated (65°C, 30 min). Following centrifugation (10,000 ×g, 5 min), 0.5 ml of the resulting supernatant was treated with genomic-binding buffer (0.5 ml), centrifuged (10,000 rpm, 5 min) and the supernatant passed through a DNA-binding column. The column was washed with Genomic wash buffer (1 ml) and 3 times with 1 ml 75% ethanol prior to dry spinning. DNA was eluted and collected from the column with 0.2 TE buffer.
The final volume of the PCR mixture was adjusted to 50 μl by adding sterile Milli Q water. Thermal cycling was performed with a TECHNE Touchgene (USA) thermal cycler. Samples were subjected to an initial denaturation step (95°C, 5 min; 95°C, 30 s) followed by annealing (40°C, 1 min). The thermal profile used was 30 cycles consisting of 1 min of primer annealing at 55°C, 1 min of extension at 72°C and 1 min of denaturation at 95°C. A final extension step consisting of 10 min at 72°C was also included. PCR products were detected by agarose gel electrophoresis (GIBCO BRL, USA) and were visualized by UV fluorescence after ethidium bromide staining. The PCR products generated by the 9F and 1492R primers were approximately 1 kb in size and were purified using the PureLink R genomic DNA kit from Invitrogen as specified by the manufacturer and sequenced bidirectionally with the primers 9F and 1492R from Sigma-Aldrich (described above). Sequencing of the final DNA extracts was done by Genius Laboratories (UK). The 16S rDNA sequences determined in this study have been deposited with Genbank.

| Phylogenetic analyses
However, 16S rDNA Sequences of UV resistant bacteria were manipulated using ClustalW for pairwise and multiple alignments.
Evolutionary analyses were conducted in MEGA6 (Tamura, Nei, & Kumar, 2004) in which the maximum composite likelihood model was used to estimate the evolutionary divergence between the sequences.
Evolutionary distance was calculated using the neighbor-joining method. Distances were obtained where the values represented the dissimilarity for each pairwise comparison (phylogenetic diversity).
A phylogenetic tree was constructed by using the neighbour-joining method and the tree reliability was determined by 2000 bootstrap replications for 95% reproducibility ( Figure 2). Accession numbers of the determined sequences are reported in Table 1.

| Short wavelength ultraviolet irradiance assays of bacteria isolates
Bacterial isolates from M. producens were streaked onto marine agar 2216 containing streptomycin sulfate antibiotic (50 μg/ml in H 2 O).
The antibiotic was added to the media upon cooling to prevent the growth of fastidious bacteria. Plate covers were aseptically removed from the plates to eliminate errors arising from UV absorption by plastics. Bacterial isolates were screened for UV resistance (302 nm, 70 W/m 2 , 0-1 h), using agar-based streak assays in petri dishes according to the method of Lin and Wang (Lin & Wang, 2001) with modifications. The coverless agar plates were turned upside down and The surface on which the plates were irradiated was sterilized with 75% ethanol between experiments. UV exposed plates and controls without UV irradiance were incubated (37°C, 48 hr) after which the numbers of colonies resisting radiation were recorded.
For UV irradiance of planktonic bacteria in broth cultures, bacteria were inoculated in marine broth 2216. The inoculum was incubated using an orbital shaker (200 rpm, 37°C, 24 hr), harvested at the stationary phase (24 hr) and the optical density at 600 nm determined.
Undiluted inoculum (500 μl) was spread onto marine agar plates for the UV irradiance assays. The populations of bacteria resisting UV irradiance exposure for 0, 15, 30, 45, and 60 min, respectively were used to determine the survival constants of the bacteria under UV irradiance.

| UV irradiance aqueous extracts assay
Bacillus licheniformis strain BLC-01 (KC660142) was cultured in marine broth media 2216 and incubated at 37°C. The broth culture, with an absorbance of 1.7 at OD 600 nm was passed through a 0.20 μm Millipore filter. The filtrate was centrifuged (3250g, 40 min) and freeze dried at −80°C to a fine powder. The powder was used to make concentrations of 2, 10, 20, and 40 mg/ml of aqueous extracts (AE) metabolites, respectively, in milli Q water. These AEs were added to cultures of the less UV resistant Marinobacterium stanieri strain MARIS-02 (KC660134) and investigated for additive UV resistance effects on the bacterium. Experiments were performed in triplicate.  Note. Sequence accession numbers of the isolated strains are shown in the second column, % similarity corresponds to previously reported organisms in the GenBank database; T in minutes represents the survival duration of the bacteria on UV exposure. The last column shows the sample site.

| Analysis of UV-irradiance data
constructed and these were based on the UV intensity, time, bandwidth, and fluence.
The survival population data of bacterial isolates upon UVB irradiation in this study suggested that UVB bacterial resistance was consistent with the following exponential decay equation: where N t and N 0 are the populations at time t and at the initial time Using the log-transformed version of Equation 3, data on surviving population N t , control population N 0 , time t and fluence H were used to estimate the survival constant of B. licheniformis exposed to UV radiation, for example but were applicable to the UV resistant bacteria in this study.

| Statistical analysis
Most data were categorical and therefore analyzed using the nonparametric test statistic Pearson's Chi-square Test for Count Data while applying continuity correction. In some cases, Fisher's exact test was used. The software R studio in R: A Language and Environment for Statistical Computing was used (Team RC, 2017).
The Two-way analysis of variance (ANOVA) was used to determine whether there were any statistically significant differences between the means of the independent groups under investigation. The list of UV resistant bacteria from this study is presented in

| Total fluence and transmittance as factors of irradiance
Cultures of surviving bacteria were observed 48 hr after incubation compared with controls of nonirradiated bacteria that grew after 24 hr. Overall, there was a marked increase in the size of colonies of (1) N t = N 0 e −KHT , (4) t = ln N 0 /(1∕kH In N t UV resistant bacteria that was not observed with controls. Total fluence H (W m −1 min −1 ) was consistent with Equation 2 and increased linearly with fluence and time as predicted. Optical densities measured as %T were critical in the total fluence reaching the film of bacteria on the plate. Dilute cultures with an absorbance of 0.1 at 600 nm and dilutions thereof did not resist UV irradiance. Only broth cultures with critical biomass replicated the UV resistance previously demonstrated by the streak method on agar plates. It was established that UV irradiation (302 nm, 70 W/m 2 ) of bacteria for 15-60 min approximates to dilution of bacteria culture in the order of 10 −7 for control populations N 0 averaging between 1.5 x 10 7 and 2.5 x 10 7 CFU.
Methanolic extracts (MEs, 0.5% v/v) in aqueous acetic acid (0.2%) did not show absorption at 334 nm indicative of MAAs, but instead they were bactericidal. All bacterial MEs showed UV absorption in the region of absorbance of DNA (260 nm).

| UV survival kinetics of the UV resistant B. licheniformis
Data for survival populations of B. licheniformis were consistent with first-order decay trends showing slopes of 0.033 (R 2 = 0.88) (Figure 3).
The survival constant k was estimated, using the equation in Figure 4.

The size of colonies of bacteria increased after UV irradiance for both
Gram-positive and Gram-negative bacteria. B. licheniformis aggregated themselves to increase their colony size during UV irradiance (see supporting data). However, Escherichia coli cells have been known to form long filaments as a result of UV irradiance (Kantor & Deering, 1966), this is the first attempt to observe increases in the sizes of colonies of bacteria subjected to UV irradiance. The survival data for B. licheniformis unequivocally showed that increases in UV exposure time reduce error margins of the survival constants (Figure 4). These results have relevance to germicidal UV irradiance of air, food, and water for sterilization. It is unclear how safe this technique is especially as UV resistant bacteria begin to grow after 48 hr.

| D ISCUSS I ON
The pantropic marine cyanobacterium M. producens is ubiquitous in the Western Indian Ocean marine ecosystem (Davies-Coleman et al.,  (Salomon et al., 2003;Tuomainen et al., 2006) but contrasts to that by Hube et al., (2009) Further, the study established that the ability to withstand UV radiation was not dependent on pigmentation. This confers with the hypothesized inability of bacteria to use photo-protective compounds. However, experiments performed on pigment-deficient mutants of bacterial isolates showed that inherent pigment conferred increased UV tolerance (Agogué, Joux, Obernosterer, & Lebaron, 2005). The variability could be due to preferential molecular targets in bacteria isolates (Nucleic acids, proteins, and lipids) by the radiation. Bacteria isolates exhibiting an exponential decay and withstanding the maximum dose of radiation with respect to time of exposure were selected and their responses compared. It has been suggested that Gram positive bacteria are better adapted to UV stress because their cell walls screen out a considerable fraction of UV radiation (Jagger, 1985 was consistent with a first-order decay curve. Similar relationships have been established for enteric bacteria in natural waters (Darakas, 2002). However, rarely any studies exist for UV medi- Marinobacterium stanieri was earlier considered to be vulnerable to UV irradiation (this study). However, it survived up to 1 hr of exposure with colonies almost exclusively located close to or against the edge of the petri dish as opposed to B. licheniformis colonies that were randomly distributed across the agar plate ( Figure 6).
Considering this phenomenon was observed only in plates that In conclusion, the study has shown that certain bacteria coexisting with M. producens have the inherent ability to resist UV radiation for prolonged times. Evolutionary factors and UV resistant molecules may be responsible for resistance. The study demonstrates the need to investigate further the real causes for prolonged UV resistance in certain Gram-positive and Gram-negative bacteria. That limited colonies of UV resistant bacteria could only be observed 48 hr after irradiation should be a cause of concern for germicidal irradiation, sterilization of air, food, and water.

ACK N OWLED G M ENTS
This work was supported through a European Union Marie Curie Burgess contributed towards the manuscript.

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest.