Correspondence: Ângela Cunha, Department of Biology & CESAM, University of Aveiro, Campus Universitário de Santiago, 3810-193 Aveiro, Portugal. Tel.: +351234370784; fax: +351234372587; e-mail: firstname.lastname@example.org
Bacteria from the surface microlayer (bacterioneuston) and underlying waters (bacterioplankton) were isolated upon exposure to UV-B radiation, and their individual UV sensitivity in terms of CFU numbers, activity (leucine and thymidine incorporation), sole-carbon source use profiles, repair potential (light-dependent and independent), and photoadaptation potential, under different physiological conditions, was compared. Colony counts were 11.5–16.2% more reduced by UV-B exposure in bacterioplankton isolates (P < 0.05). Inhibition of leucine incorporation in bacterioneuston isolates was 10.9–11.5% higher than in bacterioplankton (P < 0.05). These effects were accompanied by a shift in sole-carbon source use profiles, assessed with Biolog®EcoPlates, with a reduction in consumption of amines and amino acids and increased use of polymers, particularly in bacterioneuston isolates. Recovery under starvation was generally enhanced compared with nourished conditions, especially in bacterioneuston isolates. Overall, only insignificant increases in the induction of antibiotic resistant mutant phenotypes (RifR and NalR) were observed. In general, a potential for photoadaptation could not be detected among the tested isolates. These results indicate that UV effects on bacteria are influenced by their physiological condition and are accompanied by a shift in metabolic profiles, more significant in bacterioneuston isolates, suggesting the presence of bacterial strains adapted to high UV levels in the SML.
Among the different factors that affect bacteria, solar UV radiation (UVR) (wavelengths, 290–400 nm) could be particularly deleterious because of their simple haploid genomes with little or no functional redundancy and their small size that precludes the accumulation of protective pigments (Garcia-Pichel, 1994). The importance of microorganisms in global biogeochemical cycles and the expected increases in the exposure of aquatic organisms to damaging UV wavelengths over the course of the next decades (UNEP, 2010) makes the study of the interaction between UVR and microorganisms a timely subject.
Studies of photobiological responses of marine bacterial isolates have shown large variability in their sensitivity to UVR (e.g. Joux et al., 1999; Arrieta et al., 2000; Agogué et al., 2005; Santos et al., 2011). UV-induced changes in bacterial production, growth, survival and species composition, through the selection of phototolerant strains or by inducing photoadaptation, can, in turn, cause changes in the trophodynamics of microbial communities (Arrieta et al., 2000; Winter et al., 2001).
The toxic effects of UV-B radiation are considered to be, for the most part, the result of the absorption of photons by DNA, resulting in DNA photoproducts that block DNA replication and RNA transcription and, left unrepaired, can lead to cell death (Walker, 1984). Additionally, absorption of UVR by endogenous (e.g. porphyrins, nicotinamide coenzymes) and exogenous (e.g. humic substances and photosynthetic pigments) photosensitizers can lead to the formation of reactive oxygen species and cause oxidative damage, potentially impairing bacterial activity and viability (Baxter & Carey, 1983; Cooper, 1989; Glaeser et al., 2010).
In response to UV-induced damage, bacteria have evolved several DNA repair mechanisms that can basically be divided in dark repair and light-dependent repair. At least three dark repair mechanisms can be found in bacteria: (1) nucleotide excision repair (NER), (2) postreplication recombinational repair and (3) error-prone or mutagenic DNA repair (MDR) (Walker, 1984). The latter involves the activation of low fidelity repair polymerases, which are able to perform translesion DNA synthesis across damaged regions, at the expenses of an increased mutation rate (Goodman, 2002; Rattray & Strathern, 2003) and seems to confer an ecological advantage for microorganisms inhabiting UV-exposed habitats (Sundin & Murillo, 1999). Light-dependent repair, also known as photoenzymatic repair (Walker, 1984), uses a photolyase enzyme that can be activated by UV-A (315–340 nm) and photosynthetic active radiation (PAR) (400–700 nm) (Sancar, 1994).
Variability in UV sensitivity among bacterial strains is mainly determined by differences in intrinsic susceptibility and/or defence and repair strategies in response to damage. However, several intrinsic factors such as the nutritional state (Nystrom et al., 1992) and growth phase (Berney et al., 2006; Bucheli-Witschel et al., 2010) or external factors like temperature (Matallana-Surget et al., 2010) can also influence UV sensitivity.
The bacterioneuston inhabits the surface microlayer (SML), that is, the top millimetre of the water column, being naturally exposed to high levels of solar radiation, including the UV spectrum. Reports of enhanced prokaryote abundance at the SML in comparison with underlying waters (UW) (Agogué et al., 2004; Aller et al., 2005; Obernosterer et al., 2005; Joux et al., 2006; Santos et al., 2009) have raised the possibility of the presence of bacterial strains adapted to multi-stress conditions, including UVR, at the air–water interface. However, experimental testing of the individual sensitivity of strains isolated from the SML and UW did not confirm the different resistance to solar radiation of bacterioneuston and bacterioplankton isolates (Agogué et al., 2005). A more ecologically relevant approach to address UV resistance in natural bacterial communities can be the preliminary selection and isolation of UV-B resistant strains, likely to be relevant in the functioning of the communities under increased UV-B stress conditions (Fernández Zenoff et al., 2006).
The hypotheses of this work are as follows: (1) UV-resistant bacterioneuston and bacterioplankton isolates respond differently to UVR, in terms of damage, repair and potential for photoadaptation; (2) physiological/nutritional conditioning (starvation vs. nourishment) affects the photobiological responses of bacterioneuston and bacterioplankton isolates and (3) the pattern of photobiological responses is related to the environment of origin of the isolates.
Materials and methods
Samples from the SML and UW were collected at Ria de Aveiro (Portugal), a tidal estuary in the western coast of Portugal. Sampling was conducted on three sequential days in 2008, shortly after the summer solstice, when UV-B levels are the highest (Seckmeyer et al., 2008). Daily solar radiation doses on the period of sampling in the South of Europe range 30–35 kJ m−2 (Abboudi et al., 2008). At the time of sampling, the sky was clear and minimum wind conditions (< 2 m s−1) were observed. Triplicate samples with three sub-samples (n = 9) were collected in each sampling moment. Bacterioneuston samples were collected using a 0.25 m wide × 0.35 m Plexiglas plate, which removes the upper 60–100 μm water layer (Harvey & Burzell, 1972). Samples from underlying water were taken directly by submerging sterile glass bottles at the depth of approximately 20 cm. The physicochemical properties of the samples are provided as Supporting Information, Table S1. Water samples, kept at 4 °C and in the shade, were processed within 3 h of collection.
Kinetics of community photoinactivation and isolation of UV-resistant bacteria
Samples (30 mL) from the SML (n = 9) and from UW (n = 9) were transferred to uncovered 150-mm diameter Petri dishes and irradiated with UV-B (Philips, UV-B TL 100W/01; maximum emission peak at 311 nm, preburned for 1 h to ensure stability of light emission), with magnetic stirring, at 25 ± 0.5 °C. UV intensities were measured with a monochromator spectro-radiometer placed at the sample level (DM 300; Bentham Instruments, Reading, UK). The inactivation kinetics of whole bacterioneuston and bacterioplankton communities was assessed until an accumulated UV-B dose of 120 kJ m−2. Aliquots were taken at predetermined intervals to monitor the dose dependent variation of culturable counts during irradiation, as described later. After incubation in the dark at 25 °C for up to 7 days, colonies were counted and the log of survival was plotted as a function of the energy dose. UV-B resistant isolates were selected from the best dilution of plates corresponding to a cumulative dose of 60 kJ m−2 UV-B, which is equivalent to ambient surface UVR levels at 40–44°N latitude on sunny days near the summer solstice (Seckmeyer et al., 2008). Colonies were selected according to colour and shape and streak-plated at least three times for purification. The purity of the strains was verified by microscopic observation after Gram staining. Pure isolates were transferred to marine broth 2216 (MA 2216; Difco, Detroit, MI), and the cultures were used for subsequent molecular characterization.
Molecular characterization of the strains
Bacterial isolates were grown in marine broth on a laboratory shaker at 25 °C. Cells from stationary phase cultures in marine broth were collected by centrifugation (3200 g for 15 min), and the pellet was used for total genomic DNA extraction according to Henriques et al. (2006). BOX-PCR fingerprinting was used for molecular typing of the isolates using the BOX A1R primer (5′-CTACGGCAAGGCGACGCTGACG-3′) (Versalovic et al., 1994) as described by Rademaker et al. (2004).
For isolates displaying distinct BOX-PCR profiles, the 16S rRNA gene was amplified by PCR using primers 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-GG TTACCTTGTTACGACTT-3′) as described by Lane (1991). The PCR products were purified (Jetquick PCR Purification kit; GENOMED Gmbh, Lӧhne, Germany) and used as templates in the sequencing reactions, carried out using the primer 27F at an external laboratory (StabVida, Oeiras, Portugal). All PCR reactions were performed in an iCycler thermal cycler (Bio-Rad Laboratories, Richmond, CA) using Taq polymerase, nucleotides and buffers purchased from MBI Fermentas (Vilnius, Lithuania). Sequences were compared with sequences available in the GenBank database using the blast (Basic Local Alignment Search Tool) service to determine their approximate phylogenetic affiliations (Altschul et al., 1990). The 16S rRNA gene sequences obtained (> 430 nucleotides) along with those of related bacteria deposited in GenBank were used to construct a phylogenetic tree. The analysis was performed by the neighbour-joining method (Kimura two-parameter distance optimized criteria) using mega version 5.0 (Tamura et al., 2011). The robustness of the tree was confirmed by bootstrap analysis based on 1000 resamplings. Nucleotide sequences generated in this study have been deposited in the GenBank database under the accession numbers GQ365194–GQ365211 and GU084169–GU084170.
Preparation of the isolates for the irradiation experiments
To establish nourished conditions, bacterial isolates were grown in marine broth on an orbital shaker (100 r.p.m.) at 25 °C until late exponential phase. For starvation experiments, cells were grown in minimal nine-salt solution glucose medium (MNSS) containing 4.0 g L−1 of glucose, 2.2 g L−1 of (NH4)2SO4, and 0.54 g L−1 of K2HPO4, as described by Nystrom et al. (1992). Starvation regime was imposed by harvesting 5 mL of an exponentially growing culture by rapid filtration through a Millipore filter (pore size, 0.45 μm), washing the cells twice and resuspending in 0.2-μm-pore-size-filtered autoclaved seawater, followed by a 40-h incubation on an orbital shaker (100 r.p.m.) at 25 °C (Nystrom et al., 1992).
Cells grown under the different nutritional conditions were harvested by centrifugation (3200 g for 15 min, 20 °C), and the pellet was washed three times in 0.2-μm-pore-size-filtered, autoclaved aged seawater. Cells were resuspended in 0.2-μm-pore-size-filtered, autoclaved seawater. Bacterial abundance was determined by epifluorescence microscopy (Hobbie et al., 1977) and adjusted with 0.2-μm-pore-size-filtered autoclaved seawater to 106 cells mL−1.
Experimental testing of isolated strains
A convenient volume of cell suspension for each isolate was transferred to 150-mm diameter Petri dishes so that the depth of the liquid was < 2 mm. The lid was removed from the culture plate, and the bacterial suspension was irradiated as previously described for whole SML and UW samples with a total dose of 60 kJ m−2. Dark controls were covered with aluminium foil and included in all experiments. Each experiment was conducted with triplicate replicates and repeated in three independent moments. Culturable counts and activity (leucine and thymidine incorporation rates and sole-carbon source use profiles) were assessed as described later.
To assess cell recovery, UV-B-irradiated cell suspensions were subjected to three different treatments: (1) exposure to PAR provided by white cold lamps (Philips TLD 58W/84, total dose of 193.6 kJ m−2), (2) exposure to UV-A provided by Philips TL 100W/10R lamps (wavelength range 350–400 nm, total dose of 52.6 kJ m−2) and (3) incubation in the dark for 4 h. Aliquots of the cell suspensions were collected before and after incubation under the different recovery regimes for culturable counts and leucine and thymidine incorporation assessments.
For MDR assays, 1-mL aliquots of unirradiated and UV-B irradiated cell suspensions were added to 1 mL of 2× LB medium (Difco BD, Franklin Lakes) and grown for 12 h in total darkness with shaking. Appropriate dilutions were then plated on LB agar, and on LB agar containing either nalidixic acid (100 μg mL−1) or rifampin (75 μg mL−1) (Sigma-Aldrich, St. Louis, MO). The salinity of LB medium was adjusted to 36 practical salinity units (PSU) to correspond to that of marine broth routinely used for the maintenance of the cultures. The frequency of mutation to nalidixic acid resistance (NalR) and rifampin resistance (RifR) was calculated as the number of NalR or RifR mutants per 108 cells as described by Zhang & Sundin (2004).
To identify UV-B photoadaptive responses, the procedure described by Joux et al. (1999) was adopted. Briefly, cell suspensions were irradiated under UV-B (Philips, UV-B TL 100W/01; maximum emission peak at 311 nm) for a total dose of 60 kJ m−2. Cell suspensions were subsequently exposed to UV-A for a total dose of 52.6 kJ m−2, which was followed by a 10-h period of incubation in the dark. Cell suspensions were then subjected to a second round of UV-B radiation for a total dose of 60 kJ m−2. At every point in the procedure, aliquots of cell suspensions were removed for the determination of culturable counts.
Triplicate 100 μL aliquots were collected before and after irradiation, serially diluted in 0.2-μm-pore-size-filtered autoclaved seawater and pour-plated in marine agar 2216 (Difco, Detroit, MI). Colonies were counted after 7 days of incubation in the dark at 25 °C. The dilution and plating procedures were carried out under low-luminosity conditions to avoid photoreactivation.
Leucine and thymidine incorporation
The effects of UV irradiation on bacterial activity were estimated by the incorporation of [3H]leucine (Simon & Azam, 1989) and [3H]thymidine (Moriarty, 1986). Triplicate 1.5 mL samples and one blank [trichloracetic acid (TCA)-fixed sample] were incubated with a mixture of [3H]leucine or [3H]thymidine (Amersham; Specific Activity 64 Ci mmol−1) and the equivalent nonradioactive substrate at final saturating concentrations of 485 and 456 nM, respectively. Samples were incubated in the dark at 25 ± 0.5 °C for 1 h. Incubations were stopped by the addition of TCA to a final concentration of 5%, after which samples were centrifuged at 16 000 g for 10 min. After discarding the supernatant, 1.5 mL of 5% TCA was added and the samples were subsequently shaken vigorously on a vortex and centrifuged again. The supernatant was discarded, and 1.5 mL of UniverSol liquid scintillation cocktail (ICN Biomedicals) was added. The radioactivity incorporated in bacterial cells was measured after 3 days in a Beckman LS 6000 IC liquid scintillation counter using the external standard ratio technique (Fuhrman & Azam, 1982; Simon & Azam, 1989).
Sole-carbon source use profiles
Customized Biolog® EcoPlates containing 31 ecologically relevant C sources (six amino acids, two amines, 10 carbohydrates, seven carboxylic acids, two phenolic acids and four polymers) were used to assess changes in bacterial metabolic profiles during UV-B exposure. For operational constraints, the effect of UV exposure on sole-carbon source use profiles was only assayed under nourished conditions in representative strains of each genus retrieved (Acinetobacter sp. strain PT5I1.2G, Bacillus sp. strain PT15I3.2CB, Brevibacterium sp. strain PT5I3.3L, Micrococcus sp. strain NT25I3.2AA, Paracoccus sp. strain NT25I3.1A, Pseudomonas sp. strain NT5I1.2B, Psychrobacter sp. strain PT15I3.2CA, Sphingomonas sp. NT15I1.2B and Staphylococcus sp. strain NT25I2.1).
Biolog® EcoPlates were inoculated with 150 μL of cell suspension (unirradiated and irradiated) per well (OD595 nm = 0.2) and incubated for 72 h at 25 °C. After incubation, the optical density (OD) of the Biolog ® EcoPlate wells was measured using the microplate reader FL 600 (Bio-Tek, VE). The OD of the control well was subtracted from the OD of all the other wells to correct for background activity. Data were exported into Microsoft Excel and treated using standard software (Garland & Mills, 1991).
All experiments were repeated in three independent assays, and parameters were always determined in triplicate. Differences between treatments were assessed by one-way anova using the statistical software spss v.17. Levene test was used to assess homogeneity of variances. If variances were not homogeneous, the nonparametric Mann–Whitney test was used to assess the overall effect of treatment. Differences with P values < 0.05 were considered statistically significant.
Kinetics of community photoinactivation and isolation of UV-resistant strains
In bacterioplankton, irradiation with a 72 kJ m−2 dose resulted in the reduction in CFU concentration to the detection limit. Complete inactivation of bacterioneuston occurred with a dose of 120 kJ m−2. Exposure to 60 kJ m−2 of UV-B radiation resulted in a 1 log and 2 log reduction in CFU counts in bacterioneuston and bacterioplankton, respectively (Fig. 1).
A total of 21 distinct isolates resistant to up to 60 kJ m−2 (10 from the SML and 11 from the UW) were retrieved and identified. Isolates were affiliated with the genera Acinetobacter, Bacillus, Brevibacterium, Micrococcus, Paracoccus, Pseudomonas, Psychrobacter, Sphingomonas and Staphylococcus (Table 1). The 16S rRNA gene sequence-based phylogeny indicating the relationship of the strains isolated to previously described strains and environmental sequences is presented in Fig. 2.
Table 1. Phylogenetic affiliation, closest relative, similarity with database and microbial classification of isolated strains
In the set of isolates tested, the average reduction in CFU counts (Fig. 3a) by UV-B irradiation was significantly (P < 0.05) higher in bacterioplankton (63.9 ± 12.5%) than in bacterioneuston (52.4 ± 13.0%). Under starvation, the average reduction in colony counts was significantly (P < 0.05) higher in both bacterioplankton (76.0 ± 11.1%) and bacterioneuston (59.8 ± 13.2%), compared with nourished conditions, and also significantly higher in bacterioplankton than in bacterioneuston (P < 0.05).
Leucine and thymidine incorporation
The reduction in leucine incorporation was stronger (P < 0.05) in bacterioneuston (38.5 ± 15.8%) than in bacterioplankton (27.6 ± 10.0%) isolates. Starvation did not significantly alter (P > 0.05) the effect of irradiation on leucine incorporation neither in bacterioneuston (39.3 ± 16.6%) nor in bacterioplankton (28.7 ± 13.5%) isolates (Fig. 3b). The average inhibition of thymidine incorporation was similar in the sets of bacterioneuston (15.3 ± 9.6%) and bacterioplankton isolates (16.3 ± 4.6%) (P > 0.05) (Fig. 3c). Under starvation, inhibition of thymidine incorporation was significantly enhanced compared with nourished conditions and was higher in bacterioneuston isolates (30.2 ± 15.0%) than in bacterioplankton (22.1 ± 10.9%) (P < 0.05) (Fig. 3c). In the set of bacterioneuston isolates, the incorporation of leucine was more inhibited than the incorporation of thymidine (P < 0.05). In bacterioplankton, the effect of irradiation was similar for the two monomers.
Sole-carbon source use profiles
UV-B exposure of representative bacterial strains was also accompanied by a shift in sole-carbon source use profiles. On average, amino acids (30.3–46.9%) and polymers (12.8–34.6%) were the substrates preferred by bacterioneuston before UV exposure. After UV exposure, a significant decrease in the relative consumption rate of amino acids and amines was observed (P < 0.05), and the metabolism of bacterioneuston isolates became more dependent on polymers (up to 100%) (Fig. 4a). On average, before UV exposure, bacterioplankton consumed mostly carbohydrates (15.0–38.1%) and amino acids (24.0–35.5%). UV-B exposure did not significantly change this pattern (Fig. 4b).
Recovery under nourished conditions during postirradiation incubations, as described by CFU counts, was similar (P > 0.05) in bacterioneuston and bacterioplankton isolates with all recovery regimes. Under starvation, recovery in CFU counts in the dark regime was significantly higher in bacterioneuston (55.0 ± 16.0%) than in bacterioplankton (41.4 ± 19.3%) (P < 0.05). Compared with nourished conditions, recovery under starvation conditions was enhanced by up to 27.0% with UV-A for bacterioplankton isolates and with all recovery regimes (up to 37.9%) for the bacterioneuston isolate set (P < 0.05) (Fig. 5a–c).
The extent of recovery in leucine incorporation rate did not differ significantly (P > 0.05) between illumination regimes or isolate sets, under nourished conditions. Under starvation, bacterioneuston recovered 14.4% and 18.6% better (P < 0.05) than bacterioplankton, during the PAR and dark incubations, respectively. Compared with nourished conditions, bacterioneuston isolates recovered better under starvation conditions (up to 16.9%) (P < 0.05) in all recovery regimes tested (Fig. 5d–f).
In nourished cell suspensions, significant differences in the extent of recovery of thymidine incorporation rates between bacterioneuston and bacterioplankton isolates were only observed for the PAR regime, for which recovery was on average 24.2% higher in bacterioneuston isolates (P < 0.05). Under starvation conditions, the extent of recovery of thymidine incorporation was significantly higher in bacterioneuston than in bacterioplankton under the PAR (by 14.8%) and dark (by 21.9%) regimes (P < 0.05). In bacterioplankton, the extent of recovery of thymidine incorporation in the dark regime was 17.1% higher in nourished conditions than under starvation (P < 0.05) (Fig. 5g–i).
In general, exposure to UV-B did not result in a significant average increase in the frequency of the RifR or NalR phenotypes (P > 0.05) (Fig. 6a and b). Statistically significant differences between averaged NalR or RifR frequencies of mutation in bacterioneuston and bacterioplankton isolates and between nourished and starved conditions were not found (P > 0.05).
On average, the reduction in CFU counts during the second UV-B treatment was higher (7.1–22.0%) (P < 0.05) than during the first treatment, for both bacterioneuston and bacterioplankton sets of isolates, regardless of the nutritional condition (Fig. 7a and b).
Despite the importance of estuaries in providing crucial ecosystem functions (Mitsch & Gosselink, 2000), studies on the effects of UVR on estuarine bacterial assemblages are still scarce in the literature. In this work, UV-resistant bacteria were isolated from the SML and UW of an estuarine system, and their individual UV sensitivity was assessed in terms of CFU numbers, activity (leucine and thymidine incorporation), sole-carbon source use profiles, repair potential (light-dependent and independent) and photoadaptation potential under different physiological conditions. The fact that only a relatively small number of bacterial isolates was tested limits the ability to make generalizations for the entire natural bacterioneuston or bacterioplankton communities. However, as these were the only isolates retrieved upon irradiation of samples from the SML and UW, they probably represent the dominant culturable members of the bacterial assemblages inhabiting these compartments. Studying their UV-sensitivity responses can, therefore, provide important clues to understand how the corresponding communities might respond to enhanced UV levels.
UV resistance in bacterioneuston and bacterioplankton
Irradiation of samples from the SML and UW resulted in a steady decrease in the abundance of total culturable bacteria. Culturable survivals in bacterioneuston were still detected after an accumulated radiation dose of 108 kJ m−2, while bacterioplankton was inhibited below the detection limit with 72 kJ m−2. This could indicate the presence of bacteria with enhanced tolerance to UV-B at the sunlit SML, already reported in other light-exposed habitats, such as the plant phyllosphere (Jacobs & Sundin, 2001) and high altitude wetland waters (Fernández Zenoff et al., 2006). The bacterial isolates retrieved at different UV-B doses were affiliated to four microbial groups: the Gram-negative Gammaproteobacteria and Alphaproteobacteria and the Gram-positive Firmicutes and Actinobacteria.
The average results of the individual testing of the isolated strains revealed a lower (P < 0.05) UV-induced reduction in CFU in the bacterioneuston set, compared with bacterioplankton. The reduction in colony counts during irradiation was accompanied by a decrease in bacterial metabolic activity, assessed from the rates of leucine (protein synthesis) and thymidine (DNA synthesis) incorporation. On average, leucine incorporation was more inhibited in bacterioneuston isolates (P < 0.05), while thymidine incorporation was equally inhibited in both sets of isolates. Furthermore, in bacterioneuston isolates, leucine incorporation was significantly more inhibited than thymidine incorporation, suggesting uncoupling of DNA and protein synthesis rates upon UV-B exposure, already reported in other bacterial strains (Arrieta et al., 2000).
UV exposure was accompanied by a shift in the profile of sole-carbon sources used by bacteria. This shift was particularly significant in bacterioneuston isolates, for which a statistically significant decrease in the utilization of amines and amino acids was observed, which could indicate a metabolic restructuring to compensate for UV-induced damage to the protein synthesis apparatus (Sommaruga et al., 1997). Furthermore, in seven of the nine isolates tested, a tendency towards the use of the polymers Tween 40 and Tween 80 as a preferable carbon source was observed upon irradiation. The observed changes in metabolic profiles could result from an active transcriptional reaction of the cells to UV stress, involving a shift in metabolic strategies from growth to defence/repair, and/or could be a metabolic restructuring to compensate for UV-induced damage to sensitive enzymes. For example, in Escherichia coli, exposure to stressful conditions (temperature shift, oxidative stress and carbon starvation) induces a reduction in metabolites of the central metabolism (TCA cycle and glycolysis), as well as an increase in free amino acids, as a result of protein degradation and stalling of translation (Jozefczuk et al., 2010). Further characterization of the metabolites induced by UVR in bacteria could help to clarify the significance of the observed metabolic shift during UV-B exposure.
Recovery from UV-B induced damage and photoadaptation potential
Variability in the extent of the repair of UV-induced damage was also observed among the tested isolates. Only small, nonsignificant increases, if any, in the frequencies of NalR and RifR mutant phenotypes, used to assay the occurrence of MDR, were observed. As the expression of MDR is differentially induced depending on the amount of DNA damage experienced by the cell (Walker, 1984; Smith & Walker, 1998), the reduced induction of the mutagenic repair pathway observed in this study could indicate that bacteria are probably able to effectively repair UV-B induced DNA damage by error-free mechanisms (e.g. light-dependent repair and NER).
Whether bacterial photoadaptation occurs in aquatic environments has not been clearly established (Pakulski et al., 1998). In this work, occurrence of photoadaptation to a second round of UV exposure was not observed and the reduction in CFU numbers during the second UV exposure was, in fact, higher than in the first. Unlike DNA photoproducts, which can be effectively reversed by photoreactivation, the overproduction of ROS caused by exposure of organisms to UV-B can have far wider ranging consequences, causing oxidative lesions to biomolecules that are energetically expensive to repair (Halliwell & Gutteridge, 1999). Therefore, it is possible that after the first UV exposure, the surviving cells accumulated lesions that made them more vulnerable to the second round of irradiation, which is in accordance with our observations of accumulation of oxidative products in bacteria after UV exposure is terminated (unpublished results).
Effects of nutritional conditioning on UV-sensitivity responses
Starvation significantly enhanced UV-induced reduction in CFU counts in relation to nourished conditions. This observation suggests that starvation could enhance the sensitivity of bacteria to UV-induced damage, probably as a result of the synergistic action of energy deprivation and accumulation of reactive oxygen species during starvation (Hengge-Aronis, 2002), which is accentuated by UV exposure.
Starvation enhanced the inhibition of thymidine incorporation but had no significant effect on the response of leucine incorporation to irradiation. This indicates that UV-B inhibition of DNA synthesis, but not protein synthesis, is modulated by the nutritional status of the cell, probably due to enhanced affinity of the protein synthesis system for ATP and GTP under starvation (Jewett et al., 2009).
Nutritional conditions also affected recovery from UV-induced damage. The extent of recovery was generally higher in starved cells, particularly for bacterioneuston isolates. As starvation also induces oxidative damage (Hengge-Aronis, 2002), defence mechanisms activated in response to starvation might allow for a more efficient recovery once UV exposure is terminated. Enhanced resistance to UVR under oligotrophic conditions has been reported for Vibrio sp. strain S14 (Nystrom et al., 1992) and Enterococcus faecalis (Hartke et al., 1998).
Despite the fact that dissolved organic compounds (e.g. free amino acids) accumulate at the SML, bacterioneuston seems to display lower bacterial growth efficiencies than bacterioplankton, suggesting differences in the physiological state of bacterioneuston and bacterioplankton (Kuznetsova & Lee, 2002; Reinthaler et al., 2008). Rapidly growing cells are often more susceptible to inactivation than slower growing ones, because of the shorter time for excision repair between rounds of replication (Harm, 1980; Jagger, 1985). Whether the natural differences in physiological state of the bacterioneuston and bacterioplankton could influence their UV sensitivity and account for the enhanced UV resistance of bacterioneuston is unknown.
The photobiological responses (UV-induced damage, repair and potential for photoadaptation) of bacterial isolates retrieved from the SML and UW upon exposure to enhanced UV-B levels were studied under different nutritional conditions. Regardless of the nutritional condition, CFU counts were more affected in bacterioplankton than in bacterioneuston, while protein synthesis was more inhibited in bacterioneuston. UV-B exposure was accompanied by a metabolic shift, most notably in bacterioneuston. Recovery from UV-induced damage was enhanced under starvation conditions, particularly in bacterioneuston isolates. These observations suggest that the SML may contain a pool of bacteria adapted to cope with UV-induced stress.
This work was supported by Centre for Environmental and Marine Studies, University of Aveiro (CESAM) and the Portuguese Foundation for Science and Technology (FCT) in the form of a PhD grant to A.L. Santos (SFRH/BD/40160/2007) and a post-Doctoral grant to I. Henriques (SFRH/BPD/63487/2009). Acknowledgements are due to the two anonymous reviewers, whose insightful comments greatly improved the original manuscript.