One-year experiment on the physiological response of the Mediterranean crustose coralline alga, Lithophyllum cabiochae, to elevated pCO2 and temperature

The response of respiration, photosynthesis, and calcification to elevated pCO2 and temperature was investigated in isolation and in combination in the Mediterranean crustose coralline alga Lithophyllum cabiochae. Algae were maintained in aquaria during 1 year at near-ambient conditions of irradiance, at ambient or elevated temperature (+3°C), and at ambient (ca. 400 μatm) or elevated pCO2 (ca. 700 μatm). Respiration, photosynthesis, and net calcification showed a strong seasonal pattern following the seasonal variations of temperature and irradiance, with higher rates in summer than in winter. Respiration was unaffected by pCO2 but showed a general trend of increase at elevated temperature at all seasons, except in summer under elevated pCO2. Conversely, photosynthesis was strongly affected by pCO2 with a decline under elevated pCO2 in summer, autumn, and winter. In particular, photosynthetic efficiency was reduced under elevated pCO2. Net calcification showed different responses depending on the season. In summer, net calcification increased with rising temperature under ambient pCO2 but decreased with rising temperature under elevated pCO2. Surprisingly, the highest rates in summer were found under elevated pCO2 and ambient temperature. In autumn, winter, and spring, net calcification exhibited a positive or no response at elevated temperature but was unaffected by pCO2. The rate of calcification of L. cabiochae was thus maintained or even enhanced under increased pCO2. However, there is likely a trade-off with other physiological processes. For example, photosynthesis declines in response to increased pCO2 under ambient irradiance. The present study reports only on the physiological response of healthy specimens to ocean warming and acidification, however, these environmental changes may affect the vulnerability of coralline algae to other stresses such as pathogens and necroses that can cause major dissolution, which would have critical consequence for the sustainability of coralligenous habitats and the budgets of carbon and calcium carbonate in coastal Mediterranean ecosystems.


Introduction
Ocean acidification and climate change are currently under high scrutiny due to the threat that they represent for the biodiversity and function of marine ecosystems. Current increases in atmospheric carbon dioxide (CO 2 ) and temperature proceed at unprecedented rates in the recent history of the Earth. Atmospheric CO 2 concentration has risen from 280 ppm prior to the beginning of the industrial revolution to a current value of 388 ppm due to human activities and is expected to reach more than 700 ppm by the end of this century considering the Intergovernmental Panel on Climate Change (IPCC) scenarios (Solomon et al. 2007). Global average temperature at Earth's surface has risen by 0.7°C during the last century and is expected to rise by 3°C by 2100 (Solomon et al. 2007). Similar trends are expected for surface ocean CO 2 partial pressure (pCO 2 ) and temperature due to the oceanic uptake of anthropogenic CO 2 (Sabine et al., 2004) and to the warming of the surface mixed layer (Levitus et al. 2005).
Increasing pCO 2 in the surface ocean is likely to decrease pH by 0.2-0.4 units over the course of this century (Caldeira and Wickett 2005) which will cause major shifts in seawater chemistry, with an increase in the concentration of bicarbonate ions (HCO 3 À ) and a decrease in the concentration of carbonate ions (CO 3 2À ) and the saturation state of calcium carbonate (CaCO 3 ). Such shifts are likely to affect both calcifying and photosynthetic marine organisms due to potential changes in their physiological processes of calcification and photosynthesis that both use dissolved inorganic carbon (C T : HCO 3 À , CO 3 2À and CO 2 ) as substrate. Although the physiological response of marine organisms to ocean acidification is variable among taxa and species (Doney et al. 2009;Ries et al. 2009), the decrease in the availability of CO 3 2À is known to affect the ability of marine calcifiers to form their carbonate skeleton or shells by a decline in calcification rates. A recent meta-analysis demonstrated that calcification is generally negatively affected by ocean acidification (Kroeker et al. 2010). Seawater acidification is also likely to affect photosynthesis due to the shift in the relative proportions of CO 2 and HCO 3 À , the two forms of C T that can be used for photosynthesis. Algae can use dissolved CO 2 entering the cell by diffusion as the carbon source for photosynthesis but most of them have carbon concentrating mechanisms (CCMs) which actively take up HCO 3 À which is converted to CO 2 in the cells. This is a powerful mechanism counteracting the limited availability of CO 2 in seawater (Raven and Geider 2003). An increase in seawater pCO 2 could thus enhance photosynthesis in plants that rely exclusively on CO 2 diffusion (K€ ubler et al. 1999), while it would be less favorable to algae that use CCMs (Giordano et al. 2005). The calcareous red coralline algae coralline algae (Corallinales, Rhodophyta) are of particular interest to investigate as they conduct both photosynthesis and calcification. They are also considered among the most sensitive calcifying organisms to respond to ocean acidification due to the high solubility for their high magnesian calcite skeleton. Coralline algae are absent in naturally acidified seawater where other calcifiers can survive (Hall-Spencer et al. 2008;Martin et al. 2008). Their recruitment (Agegian 1985;Kuffner et al. 2007) and growth (Agegian 1985;Jokiel et al. 2008;Hofmann et al. 2012) are both negatively affected under elevated pCO 2 . Most recent studies show that coralline calcification is negatively affected under elevated pCO 2 (Anthony et al. 2008;Semesi et al. 2009;Gao and Zheng 2010;B€ udenbender et al. 2011;Johnson and Carpenter 2012) and that this effect is exacerbated by further ocean warming (Anthony et al. 2008). However, some authors reported a significant pCO 2 effect on calcification only in combination with increased temperature (Martin and Gattuso 2009) or a positive pCO 2 effect under moderate levels with a parabolic response of calcification in response to increased pCO 2 (Smith and Roth 1979;Ries et al. 2009). The response of photosynthesis and respiration to ocean acidification in coralline algae is poorly understood. To our knowledge, only one study has investigated the effect of increased pCO 2 on respiration in coralline algae and showed no effect (Semesi et al. 2009). The response of coralline photosynthesis to ocean acidification is variable among species but most are negatively affected (Anthony et al. 2008;Gao and Zheng 2010;Hofmann et al. 2012) with a larger impact under elevated temperature (Anthony et al. 2008). Conversely, some authors found positive (Borowitzka 1981;Semesi et al. 2009) or parabolic (Borowitzka 1981 responses. Very few studies have investigated calcification, photosynthesis, and respiration all together in coralline algae. However, these processes are tightly linked and the formation of CaCO 3 crystals in cell walls of coralline algae is suggested to be largely controlled by photosynthesis and respiration (Smith and Roth 1979;Borowitzka 1981;Gao et al. 1993;De Beer and Larkum 2001). Photosynthesis may stimulate calcification by providing an organic matrix in the cell walls where the nucleation of calcite crystals is thought to occur (Borowitzka 1981). Furthermore, photosynthesis and respiration are two processes that control pH and that may in turn influence calcification rates. Photosynthesis increases pH and thereby increases the CaCO 3 saturation state, favoring calcification (Gao et al. 1993) while respiration decreases pH and act in the opposite direction by hindering calcification (De Beer and Larkum 2001). While most recent research has focused on the response of coralline algae to ocean acidification, their response to the combined rise in pCO 2 and temperature has been poorly investigated. Marine organisms are adapted to live in specific environmental temperature ranges, and a rise in temperature is likely to have direct effects on their physiology. By making coralline algae more vulnerable to other stressors, elevated pCO 2 could have a larger impact when combined with elevated temperature than alone (Anthony et al. 2008;Martin and Gattuso 2009;Diaz-Pulido et al. 2012).
Changes in calcification and primary production in coralline algae may have profound consequences for the ecosystems that they compose from polar regions to the tropics (Johansen 1981) and in which they are a major ª 2013 The Authors. Published by Blackwell Publishing Ltd.
calcifying component of the marine benthos. Coralline algae are of particular ecological importance in shallow waters, inducing settlement and recruitment of numerous invertebrates and providing habitats for a high diversity of associated organisms (Johansen 1981). Their rigid structure contributes to the formation of numerous habitats such as rhodolith beds (Foster, 2001) or coralligenous habitats (Ballesteros 2006). Coralline algae are also of significant importance in the carbon and carbonate cycles of shallow coastal ecosystems, being major contributors to CO 2 fluxes through high community photosynthesis and respiration (Martin et al. 2005(Martin et al. , 2007 and through high CaCO 3 production and dissolution (Barron et al. 2006;Martin et al. 2007).
A better understanding of how coralline algal photosynthesis, respiration, and calcification respond to ocean acidification and warming is critical to predicting how coralline algae-based community may change in response to global environmental changes. That response may also vary depending on changes in other environmental factors and in particular irradiance which is the third major physical variable that affects both photosynthesis and calcification. The response of coralline algae to ocean acidification and/or warming has mainly been investigated through short-term (a few days to a few weeks) experimental approach, therefore neglecting the potential for physiological acclimation. In this study, we investigate, through a long-term (1 year) experiment, the combined effects of elevated pCO 2 and temperature on respiration, photosynthesis, and net calcification in the crustose coralline algae, Lithophyllum cabiochae, which is one of the main calcareous components of coralligenous communities in the Mediterranean Sea. We hypothesize that future changes in pCO 2 and temperature will incur a physiological stress in L. cabiochae thereby affecting its metabolic rates. We report on the response to elevated pCO 2 and temperature in the four seasons to assess how seasonal variations of temperature and irradiance may interact with global environmental changes.

Biological material
Specimens of the crustose coralline alga, Lithophyllum cabiochae (Boudouresque & Verlaque) Athanasiadis ( Fig. 1) were collected in the coralligenous community at ca. 25 m depth in the Bay of Villefranche (NW Mediterranean Sea, France; 43°40.73′N, 07°19.39′E) on 10 July 2006 and transported in thermostated tanks to the Villefranche Oceanography Laboratory. Algae were thoroughly cleaned of epiphytic organisms without causing any damage to the thalli. Flat thalli in the size range of 15-30 cm 2 (ca. 0.35 g dry weight cm À2 ) were selected for the experiments. The algal surface area was determined from photographs using the software Image J Version 1.37v.

Chlorophyll a analysis
Fragments of each thallus of less than 0.5 cm 2 were taken at the end of each seasonal experiment for chlorophyll a (Chl a) analyses. Thallus fragments were photographed for surface determination and immediately frozen at À80°C pending analysis. Fragments were ground in 10 mL 90% acetone with cold mortar and pestle on an ice bath in the dark. The extract was poured into 15 mL centrifuge tubes and placed in the dark at 4°C overnight. After centrifugation at 4000 rpm for 20 min, total Chl a concentration in the supernatant was determined according to the method of Strickland and Parsons (1972) using a Turner Design 10-AU fluorometer. Two successive extractions were necessary for a complete Chl a extraction.

Experimental setup
Algae were randomly assigned in four 26-L aquaria (10-11 algae per aquarium) and grown during 1 year (July 2006-July 2007) in controlled conditions of pCO 2 and temperature. A crossed (2 pCO 2 9 2 temperature levels) experimental design was set up using four independent aquaria kept at ambient (ca. 400 latm) or elevated pCO 2 (ca. 700 latm; Fig. 2a) and at ambient temperature (T, i.e. the in situ temperature that the algae experience at 25 m depth in the Bay of Villefranche) or elevated temperature (T + 3°C). There were therefore four treatments: (1) ambient pCO 2 and ambient temperature (control, labeled 400 T), (2) ambient pCO 2 and elevated temperature (400 T + 3), (3) elevated pCO 2 and ambient temperature (700 T), (4) elevated pCO 2 and elevated temperature (700 T + 3). Aquaria were continuously supplied with Mediterranean seawater at a rate of 13 L h À1 from two 110-L header tanks in which pCO 2 was adjusted by bubbling ambient air (ambient pCO 2 ) or CO 2 -enriched air (elevated pCO 2 ) obtained by mixing pure CO 2 to ambient air. Temperature was gradually changed to the desired seasonal experimental value (T = 22.0°C in summer, 17.7°C in autumn, 13.3°C in winter, and 17.7°C in spring) and maintained constant 1 month prior to physiological measurement (Fig. 2b). Ambient temperature (T) corresponded to the temperature at 25 m depth in the Bay of Villefranche. It was regularly modified according to mean changes of temperature measured between 1995 and 2006 at 20 and 30 m depth by the Service d'Observation de la Rade de Villefranche, SO-Rade, of the Observatoire Oc eanologique and the Service d'Observation en Milieu Littoral, SOMLIT/ CNRS-INSU (Fig. 2b). Temperature was controlled in each aquarium to within AE0.1°C using temperature controllers (Corema) connected to 150 W submersible heaters. Irradiance was set to the mean in situ daily irradiance at 25 m depth in the Bay of Villefranche. It was calculated from surface irradiance using attenuation coefficients measured during the experimental period. Surface irradiance (photosynthetically available radiations, PAR; in lmol photons m À2 s À1 ) was measured using a flat quantum sensor (LI-COR, LI-192SA) set on the top of the S emaphore of Saint-Jean-Cap-Ferrat, located near the sampling station. The attenuation coefficients (K PAR , mean AE SD = 0.14 AE 0.2 m À1 ) were calculated according to Kirk (1983) from irradiance profiles carried out monthly from the RV Sagitta in the Bay of Villefranche using an underwater flat quantum sensor (LI-COR, LI-192SA). The experimental irradiance was adjusted seasonally to 35, 16, 6, and 21 lmol m À2 s À1 in summer, autumn, winter, and spring, respectively, using neutral density filters (Fig. 2c). The light source consisted of two 39 W fluorescent tubes (JBL Solar Ultra Marin Day) above each aquaria. The photoperiod was adjusted weekly to the desired L:D (Light:Dark) ratio according to natural fluctuations. It varied from 9:15 in December to 15:9 in June and was maintained constant at 14:10, 10:14, 11:13, and 15:9 during the summer, autumn, winter, and spring seasonal experiments, respectively (Fig. 2d). To avoid undesirable "tank" effects, each aquarium was carefully cleaned every week and each header tank was cleaned every 3 weeks to prevent the growth of epiphytes and fouling communities or the accumulation of detritus. This maintenance and the high seawater renewal (50% h À1 ) prevented any major change in seawater composition. For more details on the experimental set up and measurements of the carbonate chemistry, see Martin and Gattuso (2009). The mean seasonal parameters of the carbonate chemistry in each aquarium are given in Table 1. Only healthy (totally pink) algae were considered for the experiment, excluding those with necroses that appeared at the end of the summer period at elevated temperature (Martin and Gattuso 2009). Five algae were selected per treatment (except in the 700 T + 3 treatment in winter and spring where n = 4). When necroses occurred, algae were replaced by healthy specimens from the remaining pool of algae in the aquaria. Algae were incubated individually in closed Perspex chambers filled with ca. 200 mL of seawater from the aquaria and continuously stirred with a magnetic stirring bar. The chambers were placed inside the aquaria in order to control temperature. Clear chambers were used to assess net production (P n ) and calcification (G) in the light while chambers with a dark plastic cover were used to assess dark respiration (R d ) and calcification (G d ). Incubations were conducted in the light at the seasonal ambient irradiance and in the dark. Additionally, in summer and winter, algae were incubated at different irradiance levels in the range of those at 25 m depth in the Bay of Villefranche ( Fig. 3) calculated from surface irradiance and using attenuation coefficients of photosynthetically available radiations (K PAR of 0.13 and 0.16 m À1 in summer and winter, respectively) as described in Martin and Gattuso (2009). The irradiance levels were adjusted using neutral density filters and controlled with a flat quantum sensor (LI-COR, LI-192SA, Li-COR Inc., Lincoln, USA). Incubations took place between 09:00 and 19:00 and were carried out for 1.5-3 h according to the season and the irradiance levels. The algae were acclimated at the desired irradiance level for at least 2 h prior to the incubation and were incubated only once a day. In the dark, algae were incubated after exposure to the ambient irradiance.
The concentration of dissolved oxygen (O 2 , lmol L À1 ) was continuously measured inside the chamber using Clark-type Strathkelvin 1302 oxygen electrodes connected to a Strathkelvin 782 oxygen meter. Water samples were taken at the beginning and at the end of the incubations for measurements of pH T (pH on the total scale) and total alkalinity (A T ) as described in Martin and Gattuso (2009). The concentration of dissolved inorganic carbon (C T ) was determined from pH T , A T , temperature and salinity using the R package seacarb (Proye and Gattuso 2003). , HCO 3 À , and dissolved inorganic carbon (C T ), and the saturation state of seawater with respect to calcite (Ω c ) and aragonite (Ω a ) are calculated from pH T , temperature, salinity, and mean seasonal A T . P n and R d expressed in terms of O 2 production and consumption, respectively (in lmol O 2 cm À2 h À1 ) were calculated as: where sO 2 is the slope of the linear regression line for change in O 2 versus time (lmol L À1 h À1 ), V is the volume of the chamber (l) and S is the surface area of the thallus (cm 2 ). Gross production (P g ) was calculated as: The changes in C T during the incubations are controlled by the metabolism of organic (photosynthesis and respiration) and inorganic carbon (calcification and dissolution). The precipitation of 1 mol of CaCO 3 decreases C T by 1 mol and A T by 2 eq according to: G and G d (lmol CaCO 3 cm À2 h À1 ) were calculated using the alkalinity anomaly technique (Smith and Key 1975) as: where DA T is the difference between initial and final A T values (leq L À1 ) and Δt is the incubation time (h). P n and R d expressed in terms of CO 2 fixation and release, respectively (in lmol C cm À2 h À1 ) were calculated as: where DC T is the difference between the initial and final C T values (lmol L À1 ). P n and G measured at different irradiance (E) levels were fitted to the P n (or G) versus E function of Platt et al. (1980) modified by the addition of a dark respiration (R d ) or calcification (G d ) term: where P s and G s are scaling parameters defined as the maximum rates of photosynthesis (or calcification) in the absence of photoinhibition (or calcification inhibition under high irradiance), a is the initial slope of the light curve, and b is the photoinhibition coefficient. The maximum rates of P g (or gross calcification) at light saturation, P g max (or G g max ) are derived as (Harrison and Platt 1986): The maximum rate of P n (or G), P n max (or G max ) are calculated as: The saturating irradiance (E k , lmol photons m À2 s À1 ) is expressed as: And the compensation irradiance (E c , lmol photons m À2 s À1 ), as:

Data analyses
The effect of pCO 2 and temperature were assessed by two-way ANOVAs and followed by Tukey HSD post hoc tests or Tukey HSD post hoc tests for unequal sample sizes (Spjotvoll/Stoline test) to separate sets of homogeneous data. When necessary, data were log-transformed to meet ANOVA requirements of normal distribution (Shapiro-Wilks test) and equality of variance (Levene test). Independent ANOVAs were performed at each season as measurements were not repeated seasonally on the same algae (replacement in case of necroses). The probability levels were adjusted for repeated analyses using a Bonferroni correction (a set to 0.05 was divided by the number of analyses). Results are expressed as mean AE standard error of the mean (SE).

Respiration
Dark respiration (R d ) presented a strong seasonal pattern following temperature variations. The highest rates were measured in summer (0.27-0.32 lmol cm À2 h À1 in terms of both O 2 consumption and CO 2 release), while the lowest rates (about threefold lower) were found in winter (0.06-0.11 lmol cm À2 h À1 ; Fig 4). Intermediate values were measured in autumn (0.09-0.20 lmol cm À2 h À1 ) and spring (0.14 to 0.17 lmol cm À2 h À1 ).
The general trend in R d for the four seasons was an increase with the 3°C rise in temperature except in summer under 700 latm where R d declined at elevated temperature. Significant main effects of temperature were detected in both autumn and winter (Table 2). R d was not affected by pCO 2 whatever the season.

Photosynthesis under ambient irradiance
Net (P n ) and gross (P g ) photosynthesis under ambient irradiance exhibited strong seasonal variations. P n was highest in summer (0.38-0.66 lmol cm À2 h À1 in terms of O 2 release or CO 2 fixation), intermediate in autumn (0.14-0.33 lmol cm À2 h À1 ) and lowest in winter and spring (0.04-0.18 lmol cm À2 h À1 ). P g also decreased from summer (0.67-0.93 lmol O 2 or CO 2 cm À2 h À1 ) to winter (0.14-0.25 lmol cm À2 h À1 ) and with intermediate values in autumn (0.34-0.43 lmol cm À2 h À1 ) and spring (0.18-0.29 lmol cm À2 h À1 ; Fig. 4). P n and P g under ambient irradiance were strongly affected by pCO 2 in summer, autumn and winter (Table 2), with a decline of 20-60% in P n and 15-30% in P g under elevated pCO 2 , relative to ambient pCO 2 . In spring, no effect of pCO 2 or temperature alone was detected and only a significant interaction between pCO 2 and temperature was observed in terms of CO 2 fluxes. No effect of temperature was found on P n and P g whatever the season, except in summer on P n in terms of CO 2 fluxes.
The content of Chl a was not affected by pCO 2 or temperature and did not show any significant difference among treatments (Table 2). It averaged 17.1 AE 0.8 lg Chl a cm À2 in summer, 17.0 AE 1.0 lg cm À2 in autumn, 18.0 AE 0.7 lg cm À2 in winter and 21.6 AE 1.1 lg cm À2 in spring.

Photosynthesis-irradiance curves
The photosynthetic response of L. cabiochae to irradiance showed different patterns in summer and winter (Fig. 5).
Maximum rate of gross photosynthesis (P g max ) was about two to threefold higher in summer than in winter, while the initial slope (a) was lower in summer than in winter. The saturating (E k ) and the compensation (E c ) irradiances were about threefold and three to sixfold higher in summer than in winter, respectively. In winter, photoinhibition was observed with a decline of photosynthesis at irradiance levels higher than 40 lmol m À2 s À1 .
In summer, the highest P g max values were observed in the 700 T treatment with values ca. 130% higher than in the control both in terms of O 2 and CO 2 fluxes (Table 3). ANOVAs revealed significant interactions between pCO 2 and temperature with contrasting responses according to the pCO 2 levels. The parameters a, E k , and E c were significantly affected by pCO 2 with lower slopes (a) and higher irradiance values (E k and E c ) under elevated pCO 2 relative to ambient pCO 2 .
In winter, P g max (both in terms of O 2 and CO 2 fluxes) was mainly affected by temperature, increasing with the 3°C rise in temperature under both ambient (140-150%) and elevated pCO 2 (110%). a was significantly affected by pCO 2 with lower slope under elevated pCO 2 relative to ambient pCO 2 . No significant effect of pCO 2 and temperature was found on E k , while E c was mainly affected by pCO 2 , being higher under elevated pCO 2 .

Calcification under ambient irradiance
Net calcification in the dark (G d ) exhibited strong seasonal changes with highest rates in summer (0.13-0.26 lmol CaCO 3 cm À2 h À1 ), intermediate in autumn (0.03-0.05 lmol cm À2 h À1 ) and lowest in winter and spring (<0.03 lmol cm À2 h À1 ). In spring, G d ranged between À0.01 and 0.00 lmol cm À2 h À1 , with negative values corresponding to a net dissolution. ANOVAs revealed that G d was significantly affected by temperature in summer but no significant difference among treatments was found at the other seasons (Table 2). Although non-significant (P = 0.07), an increase of G d with temperature can be noted under both ambient and elevated pCO 2 in winter.

Calcification-irradiance curves
The response of L. cabiochae calcification to irradiance is illustrated in Fig 5. The maximum rates of net calcification (G max ) were four to eightfold higher in summer than in winter (Table 3). a was in the same order of magnitude at both seasons, while E k decreased by a factor of 2-6 from summer to winter. Inhibition of calcification was evident in winter with a decline observed at irradiance >40 lmol m À2 s À1 under ambient temperature both in the 400 T and 700 T treatments. . Gross production, respiration, and calcification rates of Lithophyllum cabiochae in the dark and at the experimental irradiance in the four treatments (400 T, 400 T + 3, 700 T, and 700 T + 3) in summer, autumn, winter, and spring. Gross production and respiration are expressed in terms of O 2 release and CO 2 fixation (negative values for respiration correspond to O 2 consumption and CO 2 release). Data are means AE SE (n = 5, except for the 700 T + 3 treatment in winter and spring, where n = 4).  G max was significantly affected by the interaction between temperature and pCO 2 in summer and mainly affected by temperature in winter (Table 3) with an increase in G max with increasing temperature under ambient pCO 2 and a decrease or no effect under elevated pCO 2 (Table 3). E k did not differ significantly among treatments in summer but was significantly affected by pCO 2 in winter with higher values under elevated pCO 2 .
Highly significant correlations were found between G and P n in all treatments both in summer (r from 0.79 to 0.86, P < 0.001) and winter (r from 0.74 to 0.89, P < 0.001; Fig. 6).

Discussion
Response of respiration to elevated pCO 2 and temperature The changes in respiration rates observed in the present study were mainly related to temperature. The temperature dependence of respiration is well known for seaweeds (L€ uning 1990) and has already been reported for several species of coralline algae with a trend of increasing respiration with increasing temperature (see Table 4 for a review). The seasonal changes in R d in L. cabiochae are consistent with those previously reported for other species of temperate coralline algae. The threefold increase in R d between winter and summer is comparable with that observed in the free-living coralline algae (maerl), Lithothamnion corallioides by Martin et al. (2006). R d also responded positively to the 3°C rise in temperature during the colder seasons (autumn, winter, and spring) under both ambient and elevated pCO 2 but with significant effect of temperature only observed in autumn and winter. Conversely, no or negative effects of increased temperature were observed in summer. Although non-significant, a decline of R d with the 3°C rise in temperature occurred under elevated pCO 2 (700 T + 3 treatment) in summer. It may be due to an increased sensitivity to high temperature (25°C) under elevated pCO 2 leading to an early denaturation of enzymes at 25°C and a metabolic slowdown. This decline may also be the result of reduced photosynthesis in the 700 T + 3 treatment in summer and to the lower availability of photosynthates.
No significant effect of pCO 2 was detected on L. cabiochae respiration rates. This confirms recent findings that show no effect of pCO 2 on respiration in various species of soft macroalgae (Zou et al. 2011) and in crustose coralline algae (Semesi et al. 2009). Although little is known on the response of algal respiration to increased CO 2 concentrations, Zou et al. (2011) reported two possible concurrent responses: (1) the stimulation of respiration by an increase in respiratory substrates such as soluble carbohydrates due to enhanced photosynthesis and (2) a reduction in maintenance respiration due to a decrease in tissue nitrogen content (such as soluble protein and chlorophyll). However, in the present study, we found a decrease in photosynthesis with increasing pCO 2 and no change in chlorophyll content.
Response of photosynthesis to elevated pCO 2 and temperature The photosynthesis of L. cabiochae was significantly influenced by the season both in terms of production rates and photosynthetic characteristics. P g under ambient irradiance was four to sixfold higher in summer than in winter while values of P g max were two to threefold higher in summer than in winter. Such seasonal fluctuations are related to the changes of both temperature and irradiance which are fundamental parameters in the control of algal photosynthesis (L€ uning 1990). The seasonal influence of temperature and irradiance on photosynthesis has been previously reported in coralline algal species such as L. corallioides, which exhibited values of P g max twice higher in summer than in winter (Martin et al. 2006). Values of E k and E c for L. cabiochae were also considerably lower as the temperature and irradiance decreased from summer to winter. Conversely, the photosynthetic efficiency (a) was higher in winter than in summer sug- Figure 5. Net photosynthesis and calcification versus irradiance curves for Lithophyllum cabiochae in the four pCO 2 and temperature treatments (400 T, 400 T + 3, 700 T, and 700 T + 3) in summer and winter. Net photosynthesis is expressed in terms of O 2 production (negative values for respiration) and CO 2 uptake (positive values for respiration). Data are means AE SE (n = 5, except in the 700 T + 3 treatment in winter, where n = 4).
gesting a greater degree of shade acclimation in winter than in summer which is concordant with the decrease in irradiance levels between summer and winter.
In contrast with the seasonal influence of temperature on L. cabiochae photosynthesis, no effect of the 3°C warming was detected on P g under ambient irradiance. However, significant and positive effects of temperature were found on P g max in winter. This highlights the importance of examining the effect of temperature at various irradiances to assess the actual effect of warming on photosynthesis. A decline in photosynthesis can, however, occur at temperatures beyond the thermal optimum (Anthony et al. 2008). In summer, the 3°C warming effectively lead to reduced photosynthetic performance in L. cabiochae but only when combined to elevated pCO 2 . Interactive effects of pCO 2 and temperature on photosynthesis have already been reported in the tropical crustose coralline alga P. onkodes, with an exacerbated drop in net productivity under elevated temperature and pCO 2 (Anthony et al. 2008). In winter, under lower temperature P g max , P n max , and G max , maximal rate of gross production, net production and net calcification, respectively (lmol O 2 , C, and CaCO 3 cm À2 h À1 ), a, initial slope of P-E or G-E curve (lmol cm À2 h À1 (lmol photons m À2 s À1 ) À1 ), E k , light saturating point, and E c , compensation point (both in units of lmol photons m À2 s À1 ).
Values are means AE SE (n = 5 except for the 700 T + 3 treatment in winter where n = 4). P-values from the two-way ANOVAs (df = 1,16 in summer and 1,15 in winter) are shown at right. Bold type indicates Bonferroni-adjusted significance (P < 0.025). Different subscripts (a, b, and c) indicate significant difference between treatments (P < 0.025, Tukey HSD post hoc tests); nd, no difference. Transformed data are indicated: ¥ log (x). levels, the 3°C rise in temperature was beneficial for photosynthesis under both ambient and elevated pCO 2 . However, the positive effect of warming on photosynthesis was more pronounced at ambient (P g max increase of 140-150%) than at elevated pCO 2 (110%), suggesting that the combination of elevated temperature and pCO 2 may incur a physiological stress in winter or that algae did not recover their photosynthetic performance in this treatment yet.
The pCO 2 exert a strong influence on the photosynthesis of L. cabiochae. Photosynthesis under ambient irradiance was negatively affected with a decline ranging from 15 to 30% for P g and from 20 to 60% for P n at elevated pCO 2 in all seasons except in spring. Depressed photosynthesis caused by elevated pCO 2 has already been reported in the articulated coralline alga Corallina sessilis (Gao et al. 1993;Gao and Zheng 2010). A decline in growth rate under elevated pCO 2 has also been observed in several species of red algae (Israel et al. 1999;Israel and Hophy 2002). However, macroalgal species show mixed response to elevated pCO 2 . Enhancement of growth was found in the red alga Porphyra yezoensis at a pCO 2 of 1000 latm (Gao et al. 1991), while no response was reported in several species of Chlorophyta, Rhodophyta, and Phaeophyta (Israel et al. 1999;Israel and Hophy 2002). These authors attributed such non-responsiveness to the presence of CCMs which rely on HCO 3 À uptake. The ability of macroalgae to use HCO 3 À is related to its high availability in seawater relative to CO 2 . The enzyme carbonic anhydrase is involved in the CCM to convert the accumulated HCO 3 À to CO 2 for Rubisco, the enzyme that fixes CO 2 . Accordingly, the response of marine macroalgae to elevated pCO 2 depends both on the extent to which HCO 3 À is utilized. Some authors suggested that CO 2 enrichment could still result in enhanced photosynthesis even for species that can effectively use HCO 3 À because HCO 3 À utilization requires energy (Gao et al. 1991). The decrease in photosynthesis at elevated pCO 2 reported in the present study may be related to increased non-photochemical quenching and higher energy requirements under CO 2 stress as suggested by Gao and Zheng (2010) for C. sessilis. Such decline in L. cabiochae photosynthesis does not appear to be related to concomitant response in pigment content as no change in Chl a concentration was observed under elevated pCO 2 . However, other key photosynthetic pigments are involved in red algae such as phycoerythrin and phycocyanin that may decrease with increasing pCO 2 (Zou and Gao 2009). Interestingly, the response of photosynthesis to elevated pCO 2 differs when algae are exposed to various irradiance levels or to ambient irradiance. The differential response in photosynthesis with increasing pCO 2 under ambient irradiance (P g and P n at 400 latm ! 700 latm) and at saturated irradiance (P g max and P n max at 400 latm 700 latm) may result from the higher requirement of C T for photosynthesis at higher irradiance levels, especially in summer. These differences may also be related to changes in photosynthetic energy conversion efficiency (a) with increasing pCO 2 . The significant decline in a under elevated pCO 2 both in summer and winter may have slowed the metabolic process and reduce P g and P n under ambient irradiance, while it did not decrease the production capacity (P g max and P n max ) for the highest irradiance levels. Very few data are available on the effect of elevated pCO 2 on the response of photosynthesis to irradiance but a recent study of Hofmann et al. (2012) reported a similar decline of a in response to increased pCO 2 in the articulated coralline alga Corallina officinalis. In agreement with the decline in a, the E k and E c values of L. cabiochae were significantly increased under elevated pCO 2 leading to higher values of the irradiance at which photosynthesis and respiration are compensated under elevated pCO 2 . This may have major implication for photosynthesis of algae growing in dim light environments. Accordingly, L. cabiochae, which mainly experiences irradiance levels close to the ambient culture irradiance (6-35 lmol m À2 s À1 ) is likely to be physiologically disadvantaged at future CO 2 concentrations.
Response of calcification to elevated pCO 2 and temperature Like photosynthesis, the process of calcification in L. cabiochae was related to irradiance. A strong relationship is found between irradiance and calcification (G-E curves). The G-E curves followed the same trend as the P-E curves reinforcing the hypothesis that calcification and photosynthesis processes are tightly linked (Pentecost 1978). Photosynthesis effectively affects calcification through the formation of the fibrous organic matrix of the cell walls that is needed for the deposition of calcite crystals in the cell wall of coralline algae and through changes in internal pH. Changes in pH that occur in the cell wall at the site of calcification are affected by both photosynthesis and respiration so that calcification is largely regulated by these metabolic activities (Smith and Roth 1979;Gao et al. 1993;Hurd et al. 2011). Increased (or decreased) pH due to photosynthesis (or respiration) leads to increased (or decreased) concentrations of CO 3 2À and therefore can promote (or hinder) the precipitation of CaCO 3 by increasing (or decreasing) the saturation state of CaCO 3 . Marked variations in both light and dark calcification rates of L. cabiochae were observed according to seasonal changes in temperature and irradiance, with maximal rates in summer, intermediate rates in autumn, and minimal rates in winter and spring. The low values observed in spring despite increased temperature and irradiance were attributed to the poor health condition at this period (Martin and Gattuso 2009). The environmental fluctuations in temperature and irradiance are known to exert a strong control on the rate of calcification of temperate coralline algae which decreases with decreasing temperature and irradiance from summer to winter as reported for L. corallioides (Potin et al. 1990;Martin et al. 2006). In the present study, the response of L. cabiochae calcification to elevated temperature and pCO 2 differed among seasons, which shows that it is critical to take into consideration the interaction with seasonal changes of temperature and irradiance.
The 3°C rise in temperature was beneficial for L. cabiochae calcification when temperature were the lowest in autumn and winter, as already reported by Martin and Gattuso (2009) for diel net calcification when using the buoyant weight technique. At these seasons, a significant and positive effect of temperature alone was found on calcification under ambient irradiance. The calcification under ambient irradiance also increased with rising temperature in summer under ambient pCO 2 although differences were non-significant. In contrast, under elevated pCO 2 , warming was detrimental to calcification due to a significant interaction between elevated pCO 2 and temperature. Maximum rates of net calcification determined from the G-E curves in summer and winter also revealed a significant temperature effect as well as a significant interactive effect between temperature and pCO 2 with an increase in G max with increasing temperature under ambient pCO 2 but a decrease (summer) or an absence of effect (winter) under elevated pCO 2 . Such response is similar to that observed in P g max and P n max , confirming a close link between calcification and photosynthesis processes. Calcification in the dark was unaffected by increased temperature, except in summer where trend was close to that observed for light calcification with an increase in G d with increasing temperature only under ambient pCO 2 . Interestingly, the detrimental effect of warming on calcification rates only occurred under high pCO 2 . Decreasing calcification in the 700 T + 3 treatment in summer is consistent with results of Martin and Gattuso (2009) who found a higher sensitivity of L. cabiochae to warming under elevated pCO 2 . A similar response was observed in the tropical crustose coralline alga P. onkodes with a positive effect of rising temperature (+3°C) under ambient pCO 2 and a negative effect under elevated pCO 2 (Anthony et al. 2008).
The response of calcification rates to elevated pCO 2 differed according to the seasons but also to the light and dark conditions. In the dark, no effect of pCO 2 alone was found on G d at all seasons, while, in the light, the effect of pCO 2 alone was observed on G only in summer. Interestingly, comparison between treatments among temperature levels (T or T + 3) revealed that G under elevated pCO 2 were comparable or even higher (in summer) relative to G under ambient pCO 2 . In summer, G under ambient irradiance was 160% higher in the 700 T treatment than in the control. Higher values of G max (140%) were also observed in the 700 T treatment relative to the control in summer. The initial slope (a) of the G-E curve was also higher in the 700 T treatment, suggesting a higher efficiency. Interestingly, the highest rate of diel net calcification measured by Martin and Gattuso (2009) between August 9 and September 8, 2006 was also observed in the 700 T treatment. Although most of studies reported negative effects of increased pCO 2 on calcification of coralline algae (see Table 5; Gao et al. 1993;Anthony et al. 2008;Jokiel et al. 2008;Semesi et al. 2009;Gao and Zheng 2010;B€ udenbender et al. 2011;Johnson and Carpenter 2012), some studies found no (Borowitzka 1981;Martin and Gattuso 2009) or positive effects at elevated pCO 2 (Smith and Roth 1979;Ries et al. 2009). Ries et al. (2009) reported a twofold increase in net calcification at intermediate pCO 2 levels (605 and 903 latm) relative to the control in the tropical crustose coralline alga Neogoniolithon sp. while Smith and Roth (1979) reported higher rates of calcification at 1300 latm than at lower and higher pCO 2 in the articulated coralline alga Bossiella orbigniana. Such response may be related to the ability of the algae to maintain an elevated pH at the site of calcification despite reduced external pH which would facilitate CaCO 3 precipitation. Higher calcification rates of L. cabiochae in the 700 T treatment in summer corresponds to higher rate of photosynthesis (P g max and P n max ). The mitigating role of photosynthesis on calcification is further confirmed by the lack of effect of elevated pCO 2 on calcification in the dark. Changes in pH related to metabolic processes (photosynthesis or respiration) may occur in the cell walls at the site of calcification but also in the diffusion boundary layer between the algal surface and external seawater (Hurd et al. 2011). The ability of coralline algae to increase pH in their cell walls and at their surface via photosynthesis may thus increase their resilience to elevated pCO 2 . Coralline algae are also able to maintain calcification in the dark even at the relatively low pH values generated by respiration. Due to their exposure to a wide range of pH, coralline algae may have some physiological flexibility to acclimate to elevated pCO 2 . Following an earlier suggestion of Digby (1977), Hofmann et al. (2012) proposed that carbonic anhydrase may also play a role in the calcification of coralline algae by catalyzing the conversion of CO 2 into HCO 3 À and then CO 3

2À
. The stimulation of carbonic anhydrase activity could therefore also help preventing a decrease in calcification at elevated pCO 2 . However, carbonic anhydrase is also used by photosynthesis to convert HCO 3 À to CO 2 . This implies that processes of photosynthesis and calcification may thus be concurrent. The maintenance or enhancement of calcification rates under elevated pCO 2 in L. cabiochae may thus be detrimental to photosynthesis, as indicated by reduced photosynthesis under elevated pCO 2 .

Conclusion
Coralline algae are considered to be among the most vulnerable organisms to ocean acidification due to the high solubility of their high-Mg calcite skeleton. However, the present findings provide evidence of the ability of L. cabiochae to maintain or even enhance its rate of calcification under increased pCO 2 . The metabolic cost of maintaining calcification may still be detrimental to other physiological processes such as photosynthesis. We also demonstrated a particular interaction between ocean acidification and warming leading to decreased physiological rates especially during the warmer season at temperature beyond the thermal optimum. These data provide insights into the potential for physiological acclimation to future environmental changes in coralline algae. However, the present study reports only on the physiological response of healthy specimens to ocean warming and acidification, while these environmental changes may affect the vulnerability of coralline algae to other stresses such as pathogens and necroses that can cause major dissolution (Martin and Gattuso 2009), which would have major consequence for the ability of L. cabiochae population to precipitate calcium carbonate. Given the critical ecological functions of coralline algae in the coralligenous habitat, future conditions of pCO 2 and temperature in the next decades and century ahead may have major consequences for the sustainability of Mediterranean coralligenous habitats and critical ecological implications for coastal Mediterranean ecosystems.