Photochemistry and Photobiology, 2006, 82: 952-961
Symposium-in-Print: UV Effects on Aquatic and Coastal Ecosystems Ultraviolet B-photoprotection Efficiency of Mesocosm-enclosed Natural Phytoplankton Communities from Different Latitudes: Rimouski (Canada) and Ubatuba (Brazil) Bruna Mohovic*’, Sdnia M. F. Gianesella’, Isabelle Laurion2and Suzanne Roy3 ’Instituto Oceanografico, Universidade de SBo Paulo, SBo Paulo, Brazil ‘lnstitut national de la recherche scientifique, Centre Eau, Terre et Environnement, Quebec, Canada 31nstitut des Sciences de la Mer, Universite du Quebec a Rimouski, Rimouski, Canada Received 30 September 2005; accepted 17 April 2006; published online 27 April 2006 DOI: 10.1562/2005-09-30-RA-707
ABSTRACT Photoprotection against UV-B radiation (UVBR; 280-320 nm) was examined in natural phytoplankton communities from two coastal environments at different latitudes: temperate Rimouski (Canada) and tropical Ubatuba (Brazil). Mesocosm experiments were performed at these sites to examine the response of phytoplankton to increases in UVBR that corresponded to local depletions of 30% and 60% in atmospheric ozone levels (low and high UVBR treatments, respectively). A fluorescence method using a pulse amplitude modulation fluorometer (XePAM, Walz, Germany) with selective UV filters was used to estimate photoprotection, and these results were compared with an index of mycosporine-like amino acid (MAA) concentrations determined using spectrophotometry of methanol extracts. The present study provided the first evidence, to our knowledge, of the suitability of this in vivo fluorescence method for the estimation of UV photoprotection efficiency in natural phytoplankton. No significant differences were found for most of the variables analyzed between the light treatments used at both sites, but differences were found between sites throughout the duration of the experiments. Vertical mixing, used to maintain cells in suspension, likely alleviated serious UVBR-induced damage during both experiments by reducing the length of time of exposure to the highest UVBR irradiances at the surface. In Rimouski, this was the main factor minimizing the effects of treatment, because optical properties of the coastal seawater rapidly attenuated UVBR throughout the water column of the ca 2 m deep mesocosms. In this location, synthesis of MAAs and photoprotective pigments likely contributed to the observed phototolerance of phytoplankton and, hence, to their growth; however, in a comparison of the UVBR treatments, these variables showed no differences. In Ubatuba, where nutrient concentrations were significantly lower than those in Rimouski, light attenuation was less than that in Rimouski and UVBR reached the bottom of the mesocosms. UVBR penetration and the forced vertical mixing
of the cells, without the possibility of vertical migration below this photostress zone, resulted in photo-inhibition, because confinement in the mesocosms forced cells to remain constantly exposed to high levels of irradiance during the daytime. Hence, additional effects of UVBR were masked in this experiment, because cells were damaged too much and phytoplankton populations were rapidly declining. There was also an overall preservation of MAAs, in contrast with chlorophyll (Chl) degradation, in spite of the fact that this UV screening was not sufficient to counteract photo-inhibition, which suggests an important role for these molecules, either in the overall photoprotection strategy or in other physiological processes. Altogether, local water characteristics, namely attenuation, mixing, and nutrients concentration, can strongly modulate the photoprotection strategies used by natural phytoplankton populations in coastal environments.
INTRODUCTION UV radiation (UVR; 2 8 0 4 0 0 nm) is a stress factor that can affect biological and chemical processes in the aquatic environment (1,2). A consequence of the recent stratospheric ozone loss (3) is an increase in UVB radiation (UVBR; 280-320 nm). In spite of UVBR being often referred to as harmful, and UVA radiation (UVAR; 320400 nm) being referred to as beneficial because of its induction of photorepair mechanisms, any portion of UVR or even visible radiation may have either deleterious or beneficial effects on certain organisms, depending on the dose absorbed by the organism and its tolerance to it. UVR with wavelengths 2320 nm is effective in providing energy to photosynthesis (1,4,5), but only below a threshold, above which UVR decreases photosynthetic rates. Several cellular components and processes can be adversely affected by UVR, including photosynthetic rates, pigments, photosystem I1 (PSII), electron transport, adenosine-5‘triphosphate synthesis, nitrogen metabolism, protein structure and functioning, nucleic acids, cell growth and division, motility and metal complexes (1). In phytoplankton, these deleterious effects can be counteracted by several photoprotection mechanisms: (1) distribution and movements of organelles to minimize the absorption cross-sectional area (6,7), as well as movement of the
*Corresponding author ernail:
[email protected] (Bruna Mohovic) 0 2006 American Society for Photobiology 0031-8655/06
952
Photochemistry and Photobiology, 2006, 82 953 organism away from the source of radiation (8); ( 2 ) synthesis of screening substances, which absorb UVR (9-13): (3) absorption by pigments that dissipate the energy as heat or fluorescence (14,15); (4) synthesis of antioxidant enzymes (16-18) or accessory pigments that act as antioxidants (19-23) to protect the cell from reactive oxygen species: and ( 5 ) dark- and light-activated mechanisms to repair UVR-induced damage to nucleic acids (24,25). Vertical mixing alters phytoplankton exposure to UVR and can enhance or decrease photoinhibition, depending on the depth of the mixing layer (26), the mixing rate (27) and the light attenuation properties of the local seawater (28). Several studies have demonstrated that the effects of photoinhibition can also be exacerbated if cells are in nutrient-limited conditions (29), because their repair and protection mechanisms may be compromised (30). The present study was undertaken as part of a broader program t o examine the ability of the coastal phytoplankton from different latitudes to acclimate to a supplemental dose of UVBR. One of the hypotheses considered w a s that phytoplankton from tropical regions, already adapted to the high levels of U V R present at these latitudes, are better able than phytoplankton found at higher latitudes to support an additional UVBR flux. Previous studies of high-latitude environments indicated some effects of UVBR on the phytoplankton composition, community size structure, photosynthesis and production of screening substances, among other effects (26,31-34). There are still few studies in medium and low latitudes about U V B R effects on marine phytoplankton (35-38). On the basis of this hypothesis, an experimental strategy using mesocosms was established at two sites: Rimouski (Canada) and Ubatuba (Brazil). Studies by co-investigators have shown that UVBR treatments had little effect o n the biomass (chlorophyll a [Chl a] concentration) or photochemical yield (39,40) of phytoplankton sampled in the mesocosms. This lack of effect could be the result of efficient photoprotection. Hence, our specific objectives were to assess photoprotection, mainly in terms of the production of UVscreening substances, and, secondly, to address the hypothesis that tropical phytoplankton could acclimate in a short time (measured in days). Bilger et al. (41,42) developed a method to measure UV-screening substances in higher plants by evaluating the U V R transmittance of leaf epidermis using a pulse amplitude modulation (PAM) fluorometer. Their work led us to our third objective: t o use the weak signal emitted by natural phytoplankton in response to UV light t o estimate the UV-screening efficiency of these cells.
MATERIALS AND METHODS Experimental sites and setup. The first mesocosm experiment was conducted from 18 June to 27 June 2000 at the marina of Rimouski (Quebec, Canada; lat 48"30'N, long 68"29'W), in the lower St. Lawrence Estuary. The second experiment was conducted from 10 February to 16 February 2001 at the marina of Sac0 da Ribeira at Ubatuba (Sao Paul0 Brazil; Iat 23"45'S, long 45"06'W), in the Sao Paulo embayment. At both experimental sites, nine mesocosms were attached to a wharf, following an east-west orientation. The mesocosm bags were made of clear polyethylene that transmitted 85-93% of light in the 280-750 nm range. The mesocosms were filled with water collected a few kilometers offshore, one day before the beginning of the experiments. At Rimouski water was pumped from 5 m under the surface and at Ubatuba, from 8 m under the surface, using a large-intake pump (model 13D-19, Gorman-Rupp, Canada). Seawater was transported on board a ship to the experimental site (-30 min) in large, prewashed, portable outdoor pools and was transferred by gravity to a 200 L barrel fitted with 9 plastic tubes that drained into the mesocosms, ensuring simultaneous filling, homogeneity and minimal disturbance to the sampled organisms. Large zooplankton were excluded by filtering seawater through a 500 pn Nitex mesh
(Dynamic Aqua Supply Ltd., BC, Canada) that was installed over the distribution barrel. In Rimouski the total filled volume of the mesocosms was ca 2000 L and dimensions were 2.3 m in overall depth and 1.16 m in diameter. In Ubatuba the total volume was ca 1800 L and dimensions were 1.9 m in overall depth by 1.16 m in diameter. To simulate natural mixing and to maintain a homogeneous water mass, water within each mesocosm was continuously mixed by peristaltic pumps (model 2-MD-HC, Little Giant, Oklahoma City, OK) at a rate of 25 L min-' (turnover, ca 1.3 h). Water intake was located 15-28 cm below the surface, and the outflow was located at the bottom of the mesocosms. There was no addition of water to the mesocosms during the experiments. Light treatments. Both experiments consisted of three treatments, with three replicates per treatment: (1) natural ambient irradiance (N) as control; (2) natural irradiance plus low UVBR addition corresponding to 30% ozone depletion (L); and (3) natural irradiance plus high UVBR addition corresponding to 60% ozone depletion (H) (43). The H treatment corresponded approximately to the maximal UVBR enhancement if an ozone hole formed over the affected sites. UVBR enhancements in mesocosms L and H were provided, respectively, by 2 and 4 UV-B fluorescent light tubes (TUOW and 12RS, respectively: Philips, Canada) suspended 40 cm above the water surface. The fluorescent lamps, with an emission peak at 313 nm, were prebumed for at least 100 h before the beginning of the experiments to ensure light emission stability (43). The fluorescent tubes were covered with 0.13 nun cellulose diacetate sheets, which filtered the artificial UVC radiation (UVCR, 200-280 nm) emitted by the lamps. These sheets were replaced daily. To ensure equal shading effects, wooden lamp dummies were installed 40 cm over the L and N mesocosms to mimic shading of the H mesocosms caused by the lamp fixtures. In Rimouski, lamps were turned on from lOO%l500 h local time. In Ubatuba, a more sophisticated UVBR enhancement system was used. UVBR emitted by the fluorescent light tubes was adjusted using dimming ballasts (FDB4843- 120-2, Lutron Electronics, Coopersburg, PA) controlled by computer with an electronic unit linked to three UVBR detectors (model SUD240/SPS300/T/W, International Light, Newburyport, MA), each set in one mesocosm per treatment. These detectors were cross-calibrated before the experiment began. The electronic unit continuously compared readings from L and H with readings from N, providing a constant proportional enhancement of UVBR, which followed the natural variations of ambient UVBR throughout the day. Light measurements. Incident photosynthetically active radiation (PAR: 400-700 nm), UVA and UVB irradiances were recorded every minute using an IL 1700 radiometer (International Light) equipped with PAR (SUDO33/PAR/W), UVAR (SUDO33RrVAlw) and UVBR (SUD2401 SPS3OO/TlW) detectors. Irradiance throughout the water column was measured three times a day at the center of each mesocosm with a PUV-500 profiling radiometer (model 5427, Biospherical Instruments, San Diego, CA) recording at 305,313,320,340 and 380 nm and PAR wavelengths and corrected for surface irradiance using a GUV-5 10 radiometer (Biospherical Instruments) with similar sensors. The PUV radiometer provided cosinecorrected measurements of downwelling irradiance, from which we calculated the vertical attenuation coefficient (Kd, m?) and the mean depths of 1% and 50% light penetration (Zl% and Z5[,%,m) for UVBR (305 nm) and PAR. Mesocosm CUM sampling. Mesocosms were sampled daily at ca 0700 h local time at each experimental site by use of the circulating pump. For Ubatuba, only two of the three mesocosms submitted to the H treatment were considered because a technical problem occurred and the H2 bag was only filled to one-quarter of the volume of the other bags. After core sampling, volumes of 200-550 mL were filtered under reduced vacuum (700 650 >650 >710
3 1 1
RG645 R65
>650 >650
3 1
Pigment analysis. Samples from Rimouski were analyzed by hightemperature as constant inside the cuvette and similar to water temperature performance liquid chromatography (HPLC) according to ROY er a[. (461, in situ. The maximum quantum yield of PSII photochemistry (F,/Fm) was determined to establish the physiological state of the phytoplankton based on the method of Wright et al. (47). For samples from Ubatuba the For this purpose, the filter set for standard chlorophyll analyses were performed based on the method of Zapata et al. (48) b ~ ~ ~ u s populations. e fluorescence measurements was used (Table 1). the method provides enhanced peak resolution. Extraction was performed For accurate determination of the minimal fluorescence (Fo) a series of in low light with 3 mL of 95% methanol, using a tip Sonicator (%W). The 5 to 10 far-red light pulses (>700 nm, 500 ms, saturation intensity) was extract was centrifuged and the supernatant was filtered through a 0.22 p administered to extract all remaining electrons in the PS II and completely Gelman Acrodisc filter. A 50 pL volume of the filtered extract was oxidize the plastoquinone pool. Then a 600 ms saturation pulse (ca 4000 manually injected into the HPLC instrument. Chlorophylls were detected pmol photons m? s-’) was administered to measure the maximal by fluorescence (excitation, 440 nm; emission, 650 nm) and carotenoids by fluorescence (F,,). With these values, the ratio (F, - Fo)/F,, or FJF,, on-line diode-may spectrophotometry. Absorbance chromatograms were was estimated (54). To estimate the photoprotection efficiency, the read at different wavelengths (410 and 450 nm). Each peak was checked fluorescence signals Fuv, FUVBand FBGwere used to calculate FUVFBG for spectral homogeneity, and the absorption specbum was compared and FUVB/FBG ratios (where BG refers to Walz blue-green filter, Table I), as with a spectral library of standards previously created. Calibration was proposed by Bilger ef al. (41) for terrestrial plants. The principle of this done with external standards obtained commercially (DHI, Denmark) or method is that UV tight is absorbed by chlorophyll and induces prepared locally from algal cultures. The standards concentrations were fluorescence. Photosynthetic organisms capable of producing and accumudetermined through their extinction coefficients, according to Roy lating high concentrations of UV-screening compounds (such as MAAs) et al. (46). will exhibit a much lower fluorescence signal induced by UVR than will Photoprotection analysis. Spectrophotometry was used to evaluate the low UV-screened organisms, because UVR wavelengths will be absorbed UV-absorbing properties of phytoplankton. The spectral absorption was by these molecules before reaching the photosynthetic reaction centers and measured on methanol extracts of filtered cells (as in [49]) with a Perkmexciting chlorophyll. For measurements of the fluorescence signal induced Elmer Lambda 2 UV/VIS spectrophotometer equipped with an integrating by total UVR, and specifically by the UV-B band, a special UV filter set sphere (Labsphere RSA-PE-20). The extraction was done with 4 mL of pure (FS-UV, Walz) was used (Table 1). methanol on the filtration set, as in the study by Kishino et al. (50) The Statistical analysis. Statistical differences among light treatments were extracts were filtered through 0.22 pm Gelman Acrodisc filters and stored tested using repeated measures analysis of variance (ANOVA Statistica v. at -60°C in cryogenic tubes topped with argon gas. A baseline correction was performed with a methanol blank and automatically subtracted from 5.11, with days as the repeated factor. Pigment concentration and max.UV/ subsequent scans. The extracts previously stored at ambient temperature max.Ch1 a values were log,,-transformed before statistical analysis was were scanned in a Icm quartz cuvette positioned at the entrance of the inperformed to normalize variance. Sphericity assumption was checked using tegrating sphere. Scans were conducted between 280 and 750 nm at a Mauchly’s test. When significant differences were detected, Tukey’s HSD speed of 240 nm min-’ with a slit width of 1 nm and smoothing of 2 nm. post hoc test was performed to determine which treatment was different To correct for scattering losses within the measurement system, the optical from the Others. and data from each day were also by skgle-factor density (OD) at 750 nm was subtracted from all the absorbance values. ODs ANOVA. f‘ values 10.05 were accepted for significant differences, were converted to absorption coefficient values. Possible underestimation of total Concentrations of MAAs may have occurred, because extraction was brief and without sonication, and because pure methanol is not able to RESULTS extract all MAAs (51). However, the methanol extract spectra obtained provided more distinguishable peaks in the UV range when compared with Light climate the phytoplankton-specific absorption spectra obtained from filter scans (i.e. viva absorption). The absorption coefficient spectra of methanol The highest PAR, UVAR and UVBR irradiances were observed at extracts were used to determine the ratio between the maximal absorption at Ubatuba, as expected (Table 2). Kd values were higher in Rimouski the uv Peak (between 325 and 338 nm in Rimouski and between 313 and than in Ubatuba for all the analyzed bands, particularly for UVBR. 359 nm in Ubatuba) and the maximal absorption at the Chi a peak at PAR reached the bottom of the mesocosms at both sites (cf. Z1%, 666 nm, and this ratio was termed “max.UV/max.Chl a” (as in [52,53]). Fhorescence measurements. Before the performance of the analysis, Table 2), but UVAR and UVBR reached the bottom only at Water samples were dark-adapted at in situ temperature for a 60-90 min Ubatuba. UVBR (305 n m ) reached down to only approximately Period to allow relaxation of nonphotochemical quenching. A 1.4 mL one-third of mesocosm depth in Rimouski (0.92 m). A simple aliquot was then transferred to a quartz glass cuvette. A PAM fluorometer calculation was done to estimate the mean time of exposure of the fitted With a xenon lamp We-PAM, Walz, Germany) was used to measure cells to 250% of the incident UVBR. Considering that the water induced chlorophyll fluorescence daily, in the morning. Data acquisition was done With the vernier Logger Pro software. During the fluorescence column in the mesocosms at Rimouski was 2.30 m, and that the a temperame Control Unit (Walz Model US-T) maintained the turnover time of the pump was 72 min, the time that the cells
Photochemistry and Photobiology, 2006, 82 955 Table 2. Light characteristics at the two sites, representative of average conditions during the experiments. Variable Noon incident irradiance PAR (pmol photons * m-' s-') UV-A (340 nm) (pW * cm-' * nm-') UV-B (305 nm) (pW * cm-' * nm-I)
-
Rimouski
Ubatuba
1659.0 47.3 4.2
1896.0 55.8 6.65
RlMOUSKl
1.2
0.3
Kd W ' ) PAR UV-B (305 nm)
Z I (m) ~ PAR UV-B (305 nm)
0.9 5.0
0.75 2.5
5.12 0.92
6.14 1.84
1
Mesocosms depth (m) UV-B exposure per pump cycle (min) UV-B exposure per day (rnin)
3
5
4
6
7
8
9
10
UBATUBA
4 ,
ZSO% (m)
PAR UV-B (305 nm)
2
r
0.77 0.14 2.3 4.4 18.0
0.92 0.28 1.9 10.6 44.2
remained exposed in the first 0.14 m of the water column (i.e. the surface layer exposed to 250% of the incident UVBR at 305 nm) was 4.4 rnin of each 72 rnin cycle. Because the UV-B lamps were turned on for 5 h each day, the total exposure time to 250% of the applied UVBR was ca 18 min per day. According to the same reasoning, cells at Ubatuba were exposed in the first 0.28 m during 10.6 rnin per each cycle of 72 min. In Ubatuba the lights were turned on in the morning and the UVBR intensities were applied according to a percentage of the instan-
2I
T
31
9 Y
I\:
T
0 1
1
2
3
4
5
6
7
.:I I
TME (hys)
Figure 2. Nutrients (nitrate, phosphate and ammonium) mean concentration ('ISD, in pkf) for all treatments in the two sites.
taneous UV-B irradiances reaching the surface, which complicates this calculation. But if we consider that the significant UVBR increase occurred between 1000 h and 1500 h, the time of exposure to 250% of the applied UVBR will be ca 44.2 rnin per day, approximately double that observed in Rimouski.
Phytoplankton biomass and physiological conditions
Figure 1. The mean values (.+-SD)of biomass (Chl a concentration) and maximum quantum yield of PS I1 photochemistry (Fv/Fm)for all treatments in the two sites.
Because Chl a and FJF, were not significantly different among treatments (39), averages of all treatments are presented here. Phytoplankton biomass (Chl a) showed a different temporal evolution at the two sites (Fig. 1). In Rimouski Chl a concentration was low on the first few days and increased significantly after day 4, reaching ca 24 pg L-' on day 7 (bloom), and finally decreased from the day 8 until the end of the experiment. In contrast, in Ubatuba Chl a concentration was maximal at ca 4.5 pg L-' during the first 2 days and experienced a gradual decline until the end of the experiment. This difference is related to the initial levels of nutrients: the level of nitrate was >10 times higher (and the level of phosphate was >5 times higher) in the estuarine waters of Rimouski than in the tropical Ubatuba waters (Fig. 2). In Ubatuba regeneration was the only source of nutrients available, as was observed in the phosphate and ammonium evolution during the experiment. The maximum quantum yield of PS I1 photochemistry (F,P,) was low at the beginning of the experiment in Rimouski, suggesting initial photoinhibition (40) possibly caused by enclosing the natural community in relatively shallow mesocosms (Fig. 1). After that, the values increased and the cells seemed to maintain a good physiological condition after day 4. In Ubatuba, F,P, values decreased after the second day, suggesting that the physiological condition of the cells gradually deteriorated during the experiment,
956 Bruna Mohovic eta/. RlMOUSKl
0.8
UBATUBA
.
E 0.1 5
4
8
7
6
S
0.0
10
1
2
3
1
2
3
4
5
6
7
4 5 TIME (days)
6
7
3.0
?
9 2.0 I Figure 3. The mean values (+SD) of MAAs and chlorophyll a absorption (max.UV and max.Ch1 a, m-') for all treatments, and photoprotection estimated by light absorption (max.UV/max.Chl a ratio) for each of the treatments in the two sites. H* = average of High treatment replicates without the H2 mesocosm.
2z
I
l,o 0.0 4
5
8
which agrees with the decline in Chl a concentration (and in nitrate concentration, Fig. 2) observed during the experiment.
7 8 9 TIME (days)
1
0
Chl a absorption (max.Ch1 a) compared with the decrease of UV absorption (max.UV) indicates the preservation of MAAs. The concentration of the photoprotective xanthophyll cycle pigments (diatoxanthin plus diadinoxanthin) in Rimouski reached values nearly 20 times the maximum observed in Ubatuba (Fig. 4), reflecting the increase in the biomass of algae (mostly diatoms; 39). The biomass-normalized concentrations of photoprotective pigments ([diadinoxanthin diatoxanthin]/Chl a) and the extent of conversion of diadinoxanthin to diatoxanthin (diatoxanthin/[diadinoxanthin diatoxanthin]) increased with time, particularly in the latter part of the experiments. The production of p,P-carotene, an effective anti-oxidant, also increased between days 3 and 7 in the Rimouski experiment, followed by a decline (Fig. 5). In Ubatuba the biomass-normalized concentrations of xanthophyll cycle pigments and P,fi-carotene increased throughout the
Photoprotection The max.UV/max.Chl a absorption ratios obtained from methanol extracts increased during the experiments at both locations (Fig. 3). In Rimouski, the max.UV/max.Chl a absorption ratio more than doubled between days 6 and 10. Part of this increase was caused by the decrease in Chl a concentration during the same period, but there was also some accumulation of MAAs. In Ubatuba, the max.UV/max.Chl a values on the first day were approximately double those observed at the beginning of the experiment in Rimouski. The values also increased with time, but this resulted mainly from the rapid degradation of Chl a. The greater decrease in
+
+
UBATUBA
,RIMOUSKI
12
0.6
I
0.04 1
2
3
4
5
6
7
8
9
10
0.6
-
I
I
I
I
I
2
3
4
5
6
7
1
2
3
4
5
6
7
2
3 4 5 TIME (days)
6
7
0.8
rn
Eg
1
1
m 0.6
8
- 0 - H
0
0.4
t
0.2
0.2
0
0 1
2
3
4
5
6
7
8
9
10 0.3
Figure 4. The mean absolute and biomass-normalized sum (IrSD) of xanthophyll cycle pigments diadinoxanthin and diatoxanthin and the extent of conversion from diadinoxanthin to diatoxanthin for each of the treatments in the two
o,,
+ :.
..
0 1
Photochemistryand Photobiology, 2006, 82 957
Figure 5. Absolute and biomass-normalized P,P-carotene mean concentration (kSD, in nM) for each of the treatments in the two sites. H* = average of High treatment replicates without the H2 mesocosm.
experiment (Figs. 4 and 5), mainly because of the faster degradation of Chl a. The relative proportion of diatoxanthin did not vary significantly, presenting only a slight increase on day 6, for the L and H treatments (Fig. 4). The initial biomass-normalized concentration of P,P-carotene was much higher in Ubatuba than in Rimouski and it increased after day 4 (Fig. 5).
Fluorescence estimates of photoprotection The UV fluorescence ratios (FU"/FBG and FUVBFBG) decreased with time in Rimouski (Fig. 6), especially after day 2, suggesting
a gradual increase in UV-screening potential. The increase in max.UV/max.Chl a absorption ratio shows a similar pattern, although the data is available only for days 4 to 10. In Ubatuba the FW/FBGand FuvB/FBG ratios increased during the last 3 days, which indicates a decrease in UV-screening. The increase with time in the max.UV/max Chl a ratio in Ubatuba contrasts with the decrease in UV-screening assessment. The analysis of each of the two FUV/FBG ratio components showed that the strong decline in Chl a observed in Ubatuba had an effect on FBG (Fig. 7). However, there was no significant variation of Fuv.
RlMOUSKl 7.0
2 .-E (u
6.0 5.0
:m - - cFBG -
4.0 3.0
3 ii
2.0
I .o 0.0
1
1.2
ti
3
2
4
0.9 -
i-***-
10
--c. FBG
\~
e
Figure 6. The mean values (?SD) of photoprotective variables (Fuv/FBG, FUVB/FBG and max.UV/max.Chl a ) for all treatments in the two sites.
9
'.
0.3 -
t
8
j--eFuy/
0.6 -
0.0
7
6
*4=
.-E
i
5
L
-
+---..*
-\f..,* '* ,
-
-
958 Bruna Mohovic et a/. UVBR treatments effects Both for Rimouski and Ubatuba, no significant difference was found among the light treatments for any of the variables analyzed by repeated-measures ANOVA, except for the absorption ratio max.UV/max.Chl a in Ubatuba (F = 31.4; P = 0.0015).The values of this ratio were, in general, lower for the H treatment, compared to the other treatments in Ubatuba (Fig. 3). However, one-way ANOVA performed for each day of experiment did not reveal significant differences among treatments (P < 0.05). The rejection of one of the replicates (mesocosm H2) lowered the power of the statistical test. These results were not corroborated by those of the fluorescence method (FUV/FBG),which showed no difference among treatments in Ubatuba.
DISCUSSION Effects of enhanced UVBR treatments The overall lack of effects of the supplemental UVBR treatments on the measured variables was also observed for pigments composition, maximum quantum yield of PS I1 photochemistry and taxonomic composition by Roy et al. (39). They reported only a small increase in a few taxon-marker pigments, mostly during the H treatment, which indicated a slight change in community structure. As discussed by Roy et al. (39) the fact that the mesocosm-enclosed water was mixed likely alleviated serious UVBR-induced damage over the duration of the experiments, since, as shown by the calculation presented in this paper, mixing reduced the exposure time of the cells to high UVBR irradiances found at the surface. In addition, the rather strong UV attenuation in Rimouski, typical of estuarine waters, enabled the cells to avoid the UVBRexposed layer during a large portion of their vertical displacement. In a concurrent study, cells that were forced to stay at the surface for 24 h revealed significant differences between the UVBR treatments ( 5 5 ) , highlighting the beneficial effect of mixing. This was also observed in other studies using surface incubations (56,57) or when sensitive assemblages of natural phytoplankton from a deep mixed layer or recently upwelled water were trapped in a shallow mixed layer (26,58). In contrast, UV was much less attenuated in Ubatuba, resulting in UVBR penetration down to the bottom of the mesocosms. Furthermore, nutrient concentrations in Ubatuba were roughly 10 times less than in Rimouski, and mixing did not allow cells to escape from the UVBR-exposed zone. These factors likely explain the photoinhibited state of cells (low F,/Fm) and Chl a degradation in Ubatuba. The lack of response to UVBR treatments in Ubatuba’s mesocosms, also verified in two concurrent studies that used surface incubations (40,55), was probably caused mostly by the damaged physiological condition of the cells, as witnessed by the fast rate of degradation of the D1 reaction center protein of PS I1 (40). A lack of response to additional UVBR was also observed in a study where macroalgae were already inhibited by the full solar waveband (59). In natural tropical and subtropical environments, a shallow mixed layer is frequently observed (60-62), but phytoflagellates, often dominant in these communities, manage to avoid high irradiance by moving to deeper, less illuminated layers (63). A subsurface Chl maximum can then be found in fluorescence profiles, whereas primary productivity vertical profiles indicate
surface inhibition (S. M. F. Gianesella et al., unpublished results). Although mixing provided some advantage in Ubatuba by diminishing the period of exposure to the highest irradiances at the surface, it may also have been detrimental, because the small flagellates that are often found in Ubatuba’s coastal waters were unable to migrate downwards.
Photoprotective mechanisms at both locations Differences in photoprotective mechanisms were observed between the two sites. Phytoplankton from Rimouski, adapted to lower irradiances, did not rely upon high levels of MAAs, as shown by the absorption results. When trapped in a shallow mixed layer such as the mesocosm, they must depend on the vertical mixing rate and natural optical properties of seawater to minimize the effects of photoinhibition. In contrast, phytoplankton from Ubatuba can rely upon higher amounts of photoprotective substances, although these quantities of MAAs were apparently not efficient in counteracting photoinhibition. Previous studies have established that large cells are more sensitive to photosynthesis inhibition (64), whereas small cells are more sensitive to DNA damage (38), which suggests that UVBR vulnerability is size dependent. A model presented by GarciaPichel (65) suggested that sunscreens could not be used as a photoprotective mechanism of any relevance by picoplankton, and that, among nanoplankton, sunscreens could afford benefits but only at the expense of relatively heavy investments and with restricted efficiencies. It is interesting to note that, in Ubatuba, where the dominant cell size was
