Light Levels Affect Carbon Utilisation in Tropical Seagrass under Ocean Acidification

Under future ocean acidification (OA), increased availability of dissolved inorganic carbon (DIC) in seawater may enhance seagrass productivity. However, the ability to utilise additional DIC could be regulated by light availability, often reduced through land runoff. To test this, two tropical seagrass species, Cymodocea serrulata and Halodule uninervis were exposed to two DIC concentrations (447 μatm and 1077 μatm pCO2), and three light treatments (35, 100, 380 μmol m-2 s-1) for two weeks. DIC uptake mechanisms were separately examined by measuring net photosynthetic rates while subjecting C. serrulata and H. uninervis to changes in light and addition of bicarbonate (HCO3-) use inhibitors (carbonic anhydrase inhibitor, acetazolamide) and TRIS buffer (pH 8.0). We observed a strong dependence on energy driven H+-HCO3- co-transport (TRIS, which disrupts H+ extrusion) in C. serrulata under all light levels, indicating greater CO2 dependence in low light. This was confirmed when, after two weeks exposure, DIC enrichment stimulated maximum photosynthetic rates (Pmax) and efficiency (α) more in C. serrulata grown under lower light levels (36–60% increase) than for those in high light (4% increase). However, C. serrulata growth increased with both DIC enrichment and light levels. Growth, NPP and photosynthetic responses in H. uninervis increased with higher light treatments and were independent of DIC availability. Furthermore, H. uninervis was found to be more flexible in HCO3- uptake pathways. Here, light availability influenced productivity responses to DIC enrichment, via both carbon fixation and acquisition processes, highlighting the role of water quality in future responses to OA.


Introduction
Seagrass meadows are highly productive habitats that offer a wide range of ecologically and economically valuable ecosystem services [1]. Especially important is their ability to capture and convert light energy into organic matter, which then become available to other trophic levels [2,3]. This primary production is strongly determined by the amount of light available in the aquatic environment [4]. When coastal water quality declines, eutrophication and high turbidity in the water column reduce the light available to these primary producers [5]. Chronic light limitation had been shown to lower seagrass productivity and contribute to the accelerating areal loss of these habitats [6][7][8].
Seagrass productivity may also be limited by the present day seawater composition of dissolved inorganic carbon (DIC) [9,10]. However, with the rising emissions of carbon dioxide (CO 2 ) from anthropogenic activities being continually absorbed by the ocean, there is a shift in the relative proportion of each DIC species (HCO 3 -, CO 3 2-, CO 2 ) in seawater [11,12] as the pH is reduced (Ocean acidification, or OA). Importantly, the proportion of CO 2 will have the greatest percent increase (>250%, compared to 24% for HCO 3 -) among the DIC constituents under the projected pH decrease by 2100 [13]. Studies had indicated that greater availability of DIC under future OA conditions may benefit seagrasses [13][14][15]. An increase in photosynthesis and growth following exposure to increased DIC in both tropical and temperate species was previously demonstrated [16][17][18][19][20][21][22]. Moreover, since tropical and temperate seagrasses displayed a higher photosynthetic affinity for CO 2 than HCO 3 -, an increased availability of CO 2 can further boost productivity [23,24]. At natural CO 2 vent sites, growth and areal cover of seagrasses were observed to be much greater than at adjacent non-CO 2 enriched sites [15,25]. Hence, given sufficient light and nutrient availability, seagrasses could utilise the extra provisioning of DIC to enhance productivity [9,22,23,26].
Seagrasses can use bicarbonate (HCO 3 -), the dominant DIC species, for photosynthesis [27]. Bicarbonate utilization involves the dehydration and transport of HCO 3 into the plant cells [27,28]. Various HCO 3 utilization pathways that have been proposed involved the enzyme carbonic anhydrase (CA) and the presence of acidic zones maintained by proton (H + ) gradients [26,28,29]. Extracellular CA dehydrates HCO 3 to CO 2 , allowing CO 2 to diffuse into the cell. Another uptake pathway involves H + extrusion-driven co-transport of H + and HCO 3 - [30]. In general, active extrusion of H + into localized regions of the leaf boundary layer (acidic zones) for HCO 3 utilisation is energetically costly compared to passive CO 2 uptake, and thus could be limited by low light levels [22]. Light availability can influence the ability of seagrasses to exploit enriched DIC conditions for photosynthesis. Fluctuating light availability [31][32][33] could vary the carbon demand for the downstream carbon fixation cycle. On the other hand, elevated DIC (with CO 2 as substrate) increased photosynthetic efficiency and reduced light requirements of seagrasses [19,22]. So far, the interaction of light and DIC availability on seagrass photosynthesis has only been explored in a few studies [16,19,[34][35][36]. Productivity responses in seagrasses to DIC enrichment under low light conditions had been variable, ranging from no change in growth rates in Zostera marina [16], to increased growth in Amphibolis antarctica [35] and enhanced photosynthesis in Halophila ovalis, Cymodocea serrulata [34] and Thalassia hemprichii [36]. Studies demonstrating positive effects of DIC enrichment under low light [34,35] suggested that seagrass growing under low light may benefit more from DIC enrichment, through a lowered epiphyte load and/or reduced energetic demand from HCO 3 use with an increase in CO 2 supply. However, it has to be noted that some of these studies did not directly measure light levels, instead using water depth [34] and epiphyte cover [36] as proxies for light reduction. It is unclear if this DIC limitation stemmed directly from a reduced HCO 3 utilization due to low light availability, or was compounded with reduced mixing or competitive DIC uptake by epiphytic algae. This study aimed to examine the effects of light availability on carbon utilisation in two tropical seagrasses, C. serrulata and Halodule uninervis. Growth, net primary productivity and photosynthetic responses to DIC enrichment under different light treatment levels were characterised in a two-week aquaria experiment. DIC levels were chosen to represent present day and end-of-century (CO 2~1 000ppm) emission scenarios (RCP8.5) [37]. It was hypothesized that while DIC enrichment (7.5% increase relative to ambient seawater) will increase photosynthesis and growth of seagrasses, the extent of increase will be greater under lower light levels.
To further test the influence of light availability on HCO 3 use in tropical seagrasses, HCO 3 utilization mechanisms were investigated using CA and H + gradient inhibitors (CA-mediated and/or H + co-transport mediated) under contrasting light levels [23,38]. Energetic demand from using HCO 3 may make it unfavourable as a photosynthetic substrate under limiting light levels. Hence the hypothesis was that under limiting light availability, the ability to use of HCO 3 as a carbon substrate would be reduced [39,40].

Light and DIC effects on photosynthetic and growth response
Experimental set-up. The seagrasses were exposed to three different light levels (35, 100 and 380 μmol m -2 s -1 ) and two seawater DIC concentrations (high vs control; Table 1) in an indoor flow-through system over two weeks. Two weeks exposure were shown to be sufficient for inducing photosynthetic changes in the two species in a previous experiment [21]. The three light levels chosen for the experiment (35, 100 and 380 μmol m -2 s -1 ) provided 1.5, 4.3 and 16.4 mol m -2 d -1 of light per day over a 12:12 h light:dark cycle. The two lower light levels represented low light conditions that were less common, but still ecologically relevant at the source meadow. The highest light level (380 μmol m -2 s -1 ) provided an integrated daily irradiance which was similar to that most commonly encountered at Cockle Bay (12-14 mol m -2 d -1 ). The experiment was conducted at the Australian Institute of Marine Sciences, Townsville. Eighteen glass aquaria (working volume 18 litres) with three replicates for each treatment were supplied with fresh filtered (5 μm) seawater from header tanks. One pot of each seagrass species was placed into each aquarium. Leaves were cleaned of epiphytes every two days by gentle rubbing between fingers. Light and CO 2 levels were randomly assigned to each aquarium. LED lights mounted over the top of each aquarium provided illumination to cover the full sunlight spectrum (Aqua Illumination LED, USA). The LED lights were set to provide 380 μmol m -2 s -1 . Irradiance was determined using a planar irradiance meter (LICOR, USA). Light reduction was achieved by placing neutral density shade-cloth over individual aquaria. DIC concentrations were manipulated by means of feedback-regulated CO 2 input (AquaMedic, Germany) into the header tanks, as described in [42]. In all the header and aquarium tanks, diffusers and pumps were installed to ensure thorough mixing of DIC enriched water. Additional pH and temperature measurements were taken manually (pH probe: Eutech, USA; console: Oakton, USA) and pH levels calibrated to TRIS seawater standards (Batch 10, Supplied by A. Dixon, Scripps Institute of Oceanography). Salinity was measured with a handheld refractometer. Every four days, water samples were taken from each aquarium and analysed for dissolved inorganic carbon (DIC) and total alkalinity (AT) concentrations using a Vindta 3C analyser. Measured values of DIC, AT, temperature and salinity were used to calculate carbonate system parameters in USGS CO2calc software [43]. Water samples for inorganic nutrient (NH 4 + and NO 3 -) measurements were collected in duplicate, every five days, from each aquarium, filtered (0.45 μm) and analysed [44].
Growth measurements. Growth rates of seagrass shoots in the experiment were measured following [45]. At the start, all shoots were marked at the top of the bundle sheath with a needle. After two weeks of growth, the shoots were harvested; new tissue growth was excised and dried at 60°C for 48 hours before weighing. Biomass of new tissue growth was normalised to the total above-ground biomass in each pot to calculate relative growth rates (RGR, g g -1 DW day -1 ).
Photosynthesis vs irradiance curves. To characterise photosynthetic parameters, dark respiration and photosynthetic rates over a light range were measured to construct photosynthesis vs irradiance (P-E) curves two weeks after the initiation of the experiment. Oxygen consumption and evolution rates of seagrass leaves in seawater from their respective treatment aquaria (DIC enriched: 1077 ± 104 μatm; control: 447 ± 23 μatm) were monitored using optical oxygen sensors (Sensor spots-PSt3, PreSens) and a fibre-optic oxygen meter (PreSens Oxy 4) in 200 mL volume incubation chambers [21]. The chambers were incubated at 25°C water temperature in a flow-through water bath system (Lauda, Ecoline RE 106). Magnetic stirrers (2 cm × 0.8 cm, 128 revolutions per minute) provided even mixing in each chamber. One mature epiphyte-free (rank 2) leaf was held upright in the chamber to mimic natural orientation. Respiration was measured over a 20-min period in the dark. Photosynthetic rates were measured on the same leaf over a series of light steps (10,20,35,70,100,200, 380, 520, 600 μmol m -2 s -1 ), with each light step lasting 25 min. Adjustable LED lights were used to provide the different light intensities (Aqua Illumination LED). Light spectra of LED lights are provided in S2 Fig. Incubation media in chambers were replaced with fresh filtered seawater from respective aquaria prior to measurements at these light steps: 35, 100, 380 and 600 μmol m -2 s -1 . Initial trials showed that low photosynthetic rates at low light levels, and short incubation times between replacements of incubation media at higher light levels reduced the possibility of DIC limitation during incubations (S1 Table). Dark respiration and photosynthetic rates were derived by fitting a linear regression to the logged oxygen concentration data in the each chamber. Only steady-state measurements of respiration and photosynthesis (obtained after~5 min) were used for regressions. After incubation, leaves were dried (60°C, 48 h) and weighed. Rates were normalised to the dry weight of the leaf. Each optical oxygen sensor was calibrated prior to use according to method described in [46].
Net primary productivity (NPP) was taken to be the photosynthetic rate measured at the respective experimental light levels (35, 100 and 380 μmol m -2 s -1 ). To derive photosynthetic parameters, photosynthetic rates were fitted to the adapted hyperbolic tangent model of [47]. The model describes the linear increase (photosynthetic efficiency, α, mg O 2 μmol -1 photons) in photosynthetic rates with irradiance, up until the saturating irradiance (E k , μmol m -2 s -1 ) where photosynthesis plateaus at the maximum rate (P max , mg O 2 g -1 DW h -1 ). Compensation irradiance (E c , μmol m -2 s -1 ) is the light level when photosynthetic rate is equal to respiration rate.

Effect of light levels on HCO 3 uptake
To assess the effect of light on HCO 3 utilization, net photosynthesis under specific inhibition of HCO 3 uptake mechanisms was measured under two contrasting light levels. The light levels represented limiting (40 μmol m -2 s -1 ) and saturating (600 μmol m -2 s -1 ) light as determined from the previous experiment where the derived saturating light (E k ) was 100 and 98 μmol m -2 s -1 for C. serrulata and H. uninervis respectively. Therefore, the limiting light level was well below E k , and the saturating light level was above E k for both species. Bicarbonate utilization pathways can be elucidated by the effect of specific inhibitors on photosynthetic rates, as described in [28]. One pathway involves the membrane-bound extracellular CA dehydrating HCO 3 into CO 2 , which then diffuses into the cell. Addition of acetalzolamide (AZ) inhibits extracellular CA activity [28]. Another possible pathway for HCO 3 utilization consists of localised active H + extrusion to create H + gradients that facilitate the inward co-transport of H + and HCO 3 -. Addition of a buffer, e.g. Tris (hydroxymethyl) aminomethane (TRIS), dissipates the H + gradient as the buffer reacts with the extruded H + , thus altering photosynthetic rates [27]. A third HCO 3 utilization pathway consists of extracellular CA-mediated HCO 3 conversion to CO 2 in acidic zones created at H + extrusion sites. The acidic zones help concentrate CO 2 and facilitate diffusion into the cells. Strong inhibition of photosynthetic rates by the combined addition of buffer and AZ indicates a strong dependence on this pathway [29,48]. Thus, we used AZ, TRIS and their combination in the experiments described below. Net photosynthesis was determined by measuring rates of oxygen evolution in 60mL incubation chambers, similar to the procedure described above. During measurements, incubation media (fresh filtered seawater (pH 8.0) with or without inhibitor) in each chamber were stirred and maintained at 25°C. Mature and non-epiphytised leaves from non-connected shoots collected fresh from the field, were used for the measurements. Prior to the addition of the inhibitor(s), control rates of dark respiration and photosynthesis for each leaf were measured in fresh filtered seawater (pH 8.0). Dark respiration rates of the leaves were measured over 20 min, before the lights were switched on to provide 40 μmol m -2 s -1 of light. Net photosynthesis was then measured for 30 min. Subsequently, light levels were increased to 600 μmol m -2 s -1 and photosynthetic rates of the same leaves were measured. After the control measurements, incubation media for all chambers were replaced with fresh filtered seawater mixed with inhibitor solutions (AZ, TRIS or AZ+TRIS). The second set of incubations used the same leaves and followed the same sequence of measurements (in dark, at limiting light, then at saturating light). Oxygen production and consumption rates were calculated and normalised to leaf dry weight as above. For this experiment, net photosynthetic rates to TRIS and/or AZ were expressed as percentages of the control net photosynthetic rates.
Three inhibitor solutions were used: 1) 0.1 mM AZ, 2) 45 mM TRIS, and 3) combination of 0.1 mM AZ and 45 mM TRIS [38]. A stock solution of 20 mM AZ was prepared by dissolving the powder in 50 mM sodium hydroxide (NaOH). An aliquot of 350μl AZ stock solution was added to each individual chamber to achieve a final concentration of 0.1 mM. TRIS was prepared as 1 M stock solution and adjusted to pH 8.0. The buffer yielded pH 8.0 when mixed with seawater to a final concentration of 45 mM. The final TRIS concentration used here, although was lower than that commonly found in literature [22,38], was sufficient in eliciting a reduction in net photosynthetic rates.

Statistical analyses
To evaluate effects of DIC and light for each species, growth, net primary productivity and photosynthetic parameters were analysed using univariate two-factor ANOVA to test for significance in fixed effects of light (three levels) and DIC (two levels) and their interaction. Data were checked for homogeneity of variance using Levene's test and for normality using Q-Q plots. Where the assumptions of ANOVA were violated, data were square root transformed. However when the assumptions were still not met, as observed in the E c and α data for H. uninervis, the alpha-value was lowered to 0.01 to minimize the risk of a Type I error [49]. In case of significant light effects, Tukey's HSD was carried out to test which of the treatment levels were significantly different from each other. Similarly, Tukey's HSD was used to test for significance between selected groups in case of significant light and DIC interactions.
To determine if the addition of AZ and/or TRIS significantly affects net productivity of each seagrass species in fresh seawater (control), raw net productivity data were compared using paired t-tests. Further, to test if light levels affect HCO 3 utilization, paired t-tests were used to compare the change in percentage net photosynthesis between the two light levels for each inhibitor type. Arcsine transformation was not performed on the the percentage net photosynthesis data, since the data exceeded the range of 0 to 100. Instead, percentage net photosynthesis responses were square root transformed prior to analysis to meet the assumption of normality for the test. All statistical analyses were conducted using R statistical software [50].

Light and DIC effects on photosynthetic and growth response
Experimental parameters. Water temperature and salinity in the experimental aquaria did not vary considerably throughout the experiment (average ± S.D.: 23.9 ± 0.3°C and 35.9 ± 0.5 (PSS-78)). Manipulation of the CO 2 system resulted in a 140% increase in CO 2 concentration, a 10% increase in HCO 3 concentration, and a corresponding decrease of~0.3 units in pH, compared to present day control conditions (Table 1). Carbonate system parameters of the DIC treatments remained in the target range. Nutrient concentrations were similar between aquarium tanks; with an average (± S.D.) water column ammonium concentration of 0.12 ± 0.03 μM and nitrate concentration of 1.30 ± 0.55 μM. Productivity responses to light and DIC enrichment. For C. serrulata, DIC enrichment and light availability both increased relative growth rates (RGR) significantly (ANOVA: DIC p = 0.015; light p = 0.002) ( Table 2; Fig 1A). When light levels were at 100 μmol m -2 s -1 and 380 μmol m -2 s -1 , RGRs increased by 20% and 26% respectively, relative to that at 35 μmol m -2 s -1 (Tukey HSD: 100 μmol m -2 s -1 p = 0.015; 380 μmol m -2 s -1 : p = 0.002). DIC enrichment raised RGR by 13%. There was no significant interaction between light and DIC enrichment on RGR in C. serrulata. In H. uninervis, effect of DIC enrichment on RGR varied with light levels (ANOVA: p = 0.009) ( Table 2; Fig 1B). DIC enrichment increased RGR only for seagrass growing at 100 μmol m -2 s -1 (Tukey HSD: p = 0.027), but not at lower (Tukey HSD: p = 0.955) or higher (ANOVA: p = 0.905) light levels.
Photosynthetic-irradiance (P-E) curves. The adapted hyperbolic tangent model provided a good fit for all P-E curves (R 2 > 0.85; p < 0.050). Photosynthetic rates increased linearly (initial slope, α) with irradiance before plateauing off at the maximum photosynthetic rate (P max ) above saturating irradiance (E k ).
Photosynthetic efficiency (α) was significantly reduced by DIC enrichment (ANOVA: p = 0.037), but not between light levels ( Table 2). No interaction of DIC enrichment and light was detected for α in C. serrulata (Table 2; Fig 2E). Overall, DIC addition reduced α by 34% for C. serrulata. For H. uninervis, there were no main or interactive effects of light and DIC enrichment on α (ANOVA: p = 0.026; alpha lowered to 0.01) ( Table 2; Fig 2F).
In summary, both DIC enrichment and light levels influenced the response of photosynthetic parameters in C. serrulata. The change in parameters such as P max and E k depended on either DIC enrichment or light levels, and also the interaction of both factors. Conversely, photosynthesis in H. uninervis-P max and E k -seemed to be strongly influenced by light treatment but was independent of DIC enrichment.
Effects of light levels on HCO3-uptake. In C. serrulata, addition of AZ significantly reduced net photosynthesis (paired t-test: t = 4.261, df = 11, p < 0.001) (Fig 3A), indicating that activity of external CA is an important mechanism aiding in HCO 3 uptake. The reduction in net photosynthesis due to the inhibition of CA-catalysed conversion of HCO 3 -(AZ treatment) was not significantly different between light levels (paired t-test: t = 1.851, df = 5, p = 0.123) (Fig 3A). Addition of TRIS to C. serrulata also resulted in a strong reduction in net photosynthesis (paired t-test: t = 3.962, df = 11, p = 0.001) (Fig 3A). This suggests that presence of the H + gradient is essential for HCO 3 transport into the cell. Dissipation of H + gradient alone (TRIS addition) resulted in a higher reduction in net photosynthesis under saturating light (>90%), compared to that under limiting light (60%) (paired t-test: t = 10.126, df = 5, p < 0.001) (Fig 3A). When both CA and H + extrusion were inhibited (addition of both AZ and TRIS), net photosynthesis in C. serrulata was reduced significantly (paired t-test: t = 4.095, df = 11, p < 0.001) (Fig 3A). Net photosynthesis was reduced to a greater extent under saturating light than under limiting light (paired t-test: t = 7.855, df = 5, p < 0.001) (Fig 3A).

Discussion
Light availability can affect the ability of tropical seagrass to respond to increased DIC provisioning under OA. The present study examined whether C. serrulata and H. uninervis were able to adjust growth and photosynthesis responses when exposed to an enriched DIC concentration approximating predicted end-of-century pCO 2 level (1077ppm) [37] over a range of light levels. Growth of C. serrulata was stimulated by both DIC enrichment and light availability while growth of H. uninervis was strongly influenced by light availability only. Interactive effects of DIC enrichment and light treatment were evident in P-E curve parameters in C. serrulata, while photosynthetic potential in H. uninervis was affected by the experimental light treatment. The use of CA and H + gradient inhibitors highlighted important differences in carbon uptake mechanisms which may explain some of the differences in responses of the species on a physiological level.

Growth and net productivity response
The experimental light treatments represented a range of ecologically relevant light conditions that occur in the natural environment. Both C. serrulata and H. uninervis were grown in lightsaturated conditions under both moderate (100 μmol m -2 s -1 ) and high (380 μmol m -2 s -1 ) light treatments, where light levels were generally at or above E k . The only exception to this was for C. serrulata at 100 μmol m -2 s -1 under DIC enrichment, in which E k was 120 μmol m -2 s -1 . In contrast, the lowest light treatment (35 μmol m -2 s -1 ) was below E k for both species under all DIC levels, and so photosynthesis was light-limited. All light treatments were above E c , and so both seagrass species were in net carbon surplus. For C. serrulata, the effect of light and DIC enrichment on growth appeared to be additive, meaning that increasing both DIC and light increased their growth rates. Thus, the highest growth rate was observed at the high light level under DIC enrichment. Further, short term exposure to DIC enrichment meant that plants growing at 35 and 100 μmol m -2 s -1 were able to grow as fast as plants without DIC enrichment at 100 and 380 μmol m -2 s -1 respectively, such that DIC enrichment somewhat compensated for lower light availability [19]. While a previous ex-situ study showed that growth rates of C. serrulata did not respond to CO 2 enrichment [21], our current study showed an increase in growth rates for this species with DIC enrichment, albeit at a very modest 13%. Higher increases in shoot density (194-350%) and above-ground biomass (32-987%) of C. serrulata were observed in natural CO 2 vent sites with considerably greater DIC enrichment, compared to adjacent non-CO 2 enriched sites [51]. Net primary production of C. serrulata in this study appeared to be strongly limited by light, not DIC concentration. However, previous studies had indicated an increase in net primary productivity with CO 2 enrichment in this species [21,51]. Here, the results suggest that light availability plays an upstream role, relative to DIC, in the hierarchical control on seagrass photosynthesis.
Net production and growth in H. uninervis did not appear to be DIC-limited, as it was less sensitive to an increase in DIC than C. serrulata under all light treatments. However, previous work has shown that the same population of H. uninervis can respond to DIC addition by increasing net productivity and growth under similar treatment conditions [21]. Seasonal variation in carbon demand for growth and metabolism might have contributed to the observed differences in response to DIC enrichment between studies [52,53].
Both DIC enrichment and light availability had been known to stimulate productivity and growth in seagrasses [17,19,22,54]. However, while C. serrulata increased growth rates with DIC enrichment and light availability, H. uninervis did not demonstrate a growth response to DIC enrichment. The growth response of H. uninervis here was limited by light availability, consistent with its net productivity response. Growth responses to DIC enrichment can also be influenced by nutrient availability [55] and water temperature [56]. Sediment pore water nutrients were not measured in the present study and hence it was not possible to assess if overall nutrient availability was limiting seagrass productivity. Knowledge of the interactive effects of environmental factors (light, temperature, nutrients) with DIC enrichment is needed to predict future seagrass productivity responses in the field.

Photosynthetic potential
Photosynthetic response of C. serrulata to DIC enrichment depended on treatment light levels. Photosynthetic capacity (P max ) was higher in C. serrulata exposed to higher DIC levels for two weeks. C. serrulata at similar CO 2 and light conditions to those used here-DIC enrichment and at 400 μmol m -2 s -1 -did increased P max by~20% [21]. Increases in P max in response to CO 2 enrichment have also been observed in Z. marina [19] and Z. noltii [18]. A larger increase in P max was observed in plants from the limiting (60% increase) compared to the saturating light treatments (0% increase). DIC enrichment can enhance maximum photosynthetic capacity by providing more substrate for fixation and simultaneously lowering photorespiration rates [57]. Light increases maximum photosynthetic capacity by boosting the production of reducing intermediates (e.g. NADPH and ATP) for the carbon reduction cycle [58]. In theory, since both factors have independent mode of actions, their combined effect should be synergistic [59,60]. The combined sub-additive effect on P max observed in results suggested that the extent of DIC limitation was greater under low light than high light.
Saturating irradiance (E k ) in C. serrulata increased with DIC enrichment, with a greater rise in E k observed at lower light levels. Higher saturating light requirements could be driven by the higher photosynthetic capacity due to greater DIC availability, as similarly observed in Thalassia hemprichii [17]. The lowering of photosynthetic efficiency with DIC enrichment was unexpected, as increased CO 2 availability would mean less resources (i.e. ATP) were needed to procure HCO 3 for photosynthesis [39]. Under CO 2 enrichment, Zostera marina tripled the rate of light-saturated photosynthesis (i.e. P max ) to reduce the daily photoperiod required for a positive carbon balance [19]. Overall, the studies conducted so far suggest that while DIC enrichment increases the intensity of saturating irradiance required to reach maximum photosynthetic rates, it also reduces the daily period of saturating irradiance required to achieve a net carbon surplus. Photosynthesis in H. uninervis, on the other hand, was limited by light, and not by DIC availability. Photosynthetic capacity (P max ) and E k increased with increasing light levels.
Interestingly, increased DIC concentration lowered E k for H. uninervis shoots growing under low light (35 and 100 μmol m -2 s -1 ). This implies that DIC enrichment could, to a certain extent, compensate for low light levels by reducing light requirements in this species. However, the increase in productivity with light availability was higher than the increase with DIC enrichment [61], with maximum photosynthetic rates remaining the highest under high light treatment. Similarly in Z. marina, CO 2 enrichment only increased shoot production and below-ground biomass under light-replete but not light-deplete treatments [16].

Effects of light on DIC utilization
Normally the supply of CO 2 , the preferred DIC species for seagrasses [23,24], is limited by low free CO 2 concentration and diffusion rates, and slow conversion from HCO 3 to CO 2 [62].
While most seagrasses can utilise HCO 3 as a DIC source [26,27] and the concentration of total DIC is non-limiting, the high energetic cost of HCO 3 uptake makes it a less preferred substrate under low light levels [28,63]. Light fuels the generation of ATP for both carbon fixation and HCO 3 uptake [48,58]. This may explain the apparent paradox that C. serrulata in our experiment was more DIC-limited at lower light levels than at higher levels. This finding was consistent with the postulation that HCO 3 utilization is limited at lower light levels [34].
In [34], both deep water (i.e. low light) Halophila ovalis and C. serrulata showed a greater increase in relative electron transport rates (100% and 66% respectively) compared to their intertidal (i.e. high light) counterparts (30% and 20% respectively) when subjected to an 180% increase in DIC concentration. Low light availability can lower HCO 3 utilization, and OA conditions could boost DIC supply by providing more dissolved CO 2 [22,64]. Application of CA and H + gradient inhibitors showed that, in general, both CA and H + gradients are important mechanisms to allow utilisation of HCO 3 as a carbon source for the two species investigated. This was observed previously in several other seagrass species [22,28,38]. For C. serrulata, the inhibition of carbonic anhydrase (CA) and H + extrusion both reduced net photosynthesis. The strong dependence on H + extrusion for HCO 3 utilisation in C. serrulata meant that adequate light levels were needed before HCO 3 can be used efficiently for photosynthesis. Therefore, at limiting light levels this species probably depends more on CO 2 diffusion. CA-mediated HCO 3 dehydration in acidic zones was thought to be a more efficient means of HCO 3 utilization than CA conversion alone, as the CO 2 concentration at equilibrium is higher within acidic zones than at normal seawater pH, thus driving the inward diffusion of CO 2 [28]. Despite this, [26] and [38] demonstrated that extracellular CA-catalysed HCO 3 conversion to CO 2 (without H + extrusion sites) was enough to support photosynthetic demand in C. serrulata. Differences between their findings and ours could be due to long term acclimation of conspecifics to different source meadow environments, or that there were genotypic variation between seagrass populations [65,66].
Halodule uninervis appeared to vary HCO 3 uptake mechanisms, enabling it to use HCO 3 over a wider light range for photosynthesis. Under low light conditions, CA-mediated conversion of HCO 3 contributed substantially to the carbon supply for photosynthesis [28]; when light became saturating, HCO 3 uptake was supplemented by H + co-transport with HCO 3 -.
Extracellular CA-mediation of HCO 3 uptake is less likely to depend on light availability, when compared to the H + extrusion-driven co-transport of HCO 3 - [26,29]. This flexibility between HCO 3 utilization pathways suggested H. uninervis was able to mediate, to a certain extent, DIC limitation under low light conditions. Similarly, efficient use of HCO 3 utilization mechanisms was thought to account for the lack of photosynthetic responses to DIC enrichment in marine macroalgae [67].
Differential sensitivity to photosynthetic carbon between seagrass species could be due to species variation in DIC utilization mechanisms [38,64] and extent of carbon-limitation [9,64]. Species such C. serrulata (this study) and Thalassia testudinum [64] would benefit more from increased CO 2 supply under OA conditions since they were limited in their ability to utilise the dominant HCO 3 in seawater. Less DIC limited species, like H. uninervis (this study), Halodule wrightii and Syringodium filiforme might show a smaller response to DIC enrichment [64].

Light and OA as drivers of seagrass productivity
Under future scenarios of OA, marine macrophytes like seagrass could benefit, from both increased DIC concentration and a proportional rise in CO 2 [15,25]. While short-term (e.g. this study) and long-term [68] studies have documented varying degrees of DIC limitation in seagrasses, physiological processes that could regulate responses to increased DIC over intermediate time-scales remain unexplored for seagrasses. Epiphytic algae may also benefit from higher CO 2 [35], and in turn compete with seagrasses for the same resources. Their effects would vary with their composition, biomass and the turnover rates of seagrass leaves [69]. Furthermore, many seagrass habitats are primarily light-limited [8]. The range of light levels used in this experiment, representing the recent light exposure history (~five years) of the seagrasses at their source meadow, was already much reduced compared to pre-European settlement due to a four-times increase in sediment runoff [70]. Reduced light availability, due to increased sedimentation and epiphytic algal growth, can negate positive OA effects on seagrass growth [35]. Our study showed that the rise in light availability elicited a greater increase in seagrass productivity than DIC enrichment. Compared to OA, variation in water clarity occurs over shorter frequencies and with greater intensities, thus playing a more imminent role in controlling seagrass productivity. Hence, to ensure continued productivity in seagrass meadows in the future, changes in water quality and OA has be to studied and managed in unison. At 100% blue channel only (blue), 100% white channel only (orange) and 100% both blue and white channels (black). Light spectra were measured using Jazz spectrometer on SpectraSuite software (Ocea-nOptics, USA). (JPG) S1 Table. Average incubation time, and approximate time range for DIC limitation to occur in 200 mL of ambient seawater during trials, for each light step of P-E curve. Leaf material used during trials was 5-12 mg in dry weight. Light steps in bold indicate a replacement of fresh filtered seawater prior to incubation during actual measurements.

Supporting Information
Rath provided valuable field assistance. Comments from anonymous reviewers improved the manuscript.