We grew a non-bicarbonate using red seaweed, Lomentaria articulata (Huds.) Lyngb., in media aerated with four O2 concentrations between 10 and 200% of current ambient [O2] and four CO2 concentrations between 67 and 500% of current ambient [CO2], in a factorial design, to determine the effects of gas composition on growth and physiology. The relative growth rate of L. articulata increased with increasing [CO2] up to 200% of current ambient [CO2] but was unaffected by [O2]. The relative growth enhancement, on a carbon basis, was 52% with a doubling of [CO2] but fell to 23% under 5× ambient [CO2]. Plants collected in winter responded more extremely to [CO2] than did plants collected in the summer, although the overall pattern was the same. Discrimination between stable carbon isotopes (Δ13C) increased with increasing [CO2] as would be expected for diffusive CO2 acquisition. Tissue C and N were inversely related to [CO2]. Growth in terms of biomass appeared to be limited by conversion of photosynthate to new biomass rather than simply by diffusion of CO2, suggesting that non-bicarbonate-using macroalgae, such as L. articulata, may not be directly analogous to C3 higher plants in terms of their responses to changing gas composition.
Atmospheric concentrations of CO2 and O2 have changed dramatically over the evolutionary lifetime of macroalgae as a result of natural variations in ocean/atmosphere dynamics ( Beerling 1994; Raven et al. 1994 ). The recent extreme changes in atmospheric [CO2] caused by anthropogenic processes are unprecedented, not in terms of amplitude, but in rapidity, with a doubling of atmospheric CO2 concentration predicted in the order of decades. Localized variation in [CO2] and [O2] can be even more extreme and rapid when gaseous exchange and equilibrium with the rest of the environment is restricted, for example in tidal pools at low tide, or to a lesser extent in submerged, dense algal turfs where flow is inhibited. The effects of changing atmospheric gas composition on benthic macroalgae, the phototrophs which dominate such environments, are largely unknown.
Most marine macrophytes have mechanisms which allow them to utilize the bicarbonate (HCO3−) of dissolved inorganic carbon (DIC) as a source of CO2 for Rubisco ( Sand-Jensen & Gordon 1984; Maberly 1990; Johnston 1991). The high concentration of HCO3− in normal seawater results in saturation or near saturation of carbon fixation with DIC in species which have such mechanisms ( Beer & Koch 1996). However, under some conditions in the natural environment, DIC can limit photosynthesis of macroalgae ( Holbrook et al. 1988 ; Levavasseur et al. 1991 ) and benthic microalgae ( Glud, Ramsing & Revsbech 1992). A few green and red macroalgae are unable to make either direct or indirect use of HCO3− ( Maberly 1990; Raven et al. 1995 ), or to acquire exogenous dissolved CO2 by active transport as has been shown for a freshwater green alga ( Rotatore, Lew & Colman 1992) and hypothesized for numerous other phototrophs ( Raven 1997a). The ‘non-bicarbonate-using’ species presumably acquire DIC by diffusion of dissolved CO2 and have no carbon-concentrating mechanism (CCM). Species which take up dissolved CO2 solely by diffusion are likely to be carbon-limited under most conditions of light and nutrient supply in the intertidal and shallow subtidal zones, although some subtidal species may not receive enough light to be DIC-limited ( Maberly, Johnston & Raven 1992). We hypothesized that changing [CO2] could have significant effects on the productivity of macroalgae, particularly those which have no CCM. One such organism is Lomentaria articulata which lives in dense turfs and shaded microhabitats in the intertidal zone ( Hiscock 1986). The apparent restriction of L. articulata to low light microhabitats may be related to its inability to use higher photon flux densities in the absence of a high internal DIC supply ( Johnston, Maberly & Raven 1992; Kübler & Raven 1994). Therefore, it seems possible that the changing [CO2] may affect not only the potential productivity of this non-bicarbonate-using macroalgal species but also its potential niche in the intertidal zone.
Increasing atmospheric [CO2] affects the productivity of terrestrial plants to varying degrees, depending on their mechanism of initial C fixation. Reviews of the large number of studies on the effects of increased atmospheric [CO2] on primary productivity have concluded that, on average, a doubling of atmospheric [CO2] results in an enhancement of growth rates of C3 plants by 40% and of C4 plants by 20% ( Poorter 1993; Tissue et al. 1995 ; Lloyd & Farquhar 1996; Ghannoun et al. 1997 ; LeCain & Morgan 1998). The differential response to [CO2] has led to suggestions that increasing atmospheric [CO2] may lead to shifts in species composition in terrestrial systems ( Wray & Strain 1987; Bazzaz & Garbutt 1988; Curtis et al. 1989 ). Benthic macrophyte communities are frequently dominated by macroalgal species which are believed to be DIC-saturated for growth via CCMs and are therefore expected to be insensitive to increased [CO2]. However, this expectation has never been thoroughly tested for growth, as opposed to short-term photosynthesis ( Raven 1997a). Similarly, although the growth rates of seagrasses and non-bicarbonate-using macroalgae are in general not saturated by current [CO2] ( Johnston et al. 1992 ; Beer & Koch 1996), the possibility that their growth rates would be affected by changing atmospheric gas concentrations has not been previously tested.
Here, we simultaneously test the hypotheses that (1) changing [CO2] and [O2] in the gas phase can affect the growth rate of Lomentaria articulata, and (2) the growth response to changing [CO2] is determined by the diffusive mechanism of inorganic carbon acquisition in this species. We use [CO2] and [O2] both above and below the current atmospheric concentrations in ranges wider than those estimated to have occurred in recent evolutionary history ( Beerling 1994; Raven et al. 1994 ) as well as those projected for the immediate future, by growing plants in media aerated with 16 combinations of [CO2] and [O2]. We use a range of [O2] as well as [CO2] to gain insight into the mechanism of inorganic carbon acquisition in this species. Terrestrial C3 plants show the expected (from diffusive gas exchange and C3 biochemistry) growth inhibition by [O2] in competition with [CO2], while growth of terrestrial C4 plants is oxygen-insensitive at up to twice the present atmospheric level ( Raven et al. 1994 ). Physiologically significant [O2] changes are not, of course, expected (or measured) as a result of anthropogenic [CO2] changes ( Keeling, Piper & Heimann 1996).
The work of Brooks & Farquhar (1985) predicts that the ratio of oxygenase to carboxylase activity of Rubisco would be low (≈ 5%) at the typical growth temperature of Lomentaria articulata (10 °C) under ambient gas composition, assuming a selectivity factor appropriate for Spinacia. The limited data available for selectivity factors of Rubisco from red algae (with CCMs) indicate a specificity 1·5–2 times that typical of terrestrial C3 plants ( Read & Tabita 1994; Uemura et al. 1996 ) so the proportion of photorespiratory inhibition of photosynthesis for a given atmospheric gas composition might be only 0·5–0·67 of that expected for Spinacia. The ratio of [O2] to [CO2] would alter the expected photorespiratory inhibition of photosynthesis in a typical C3 plant (i.e. doubling [O2] would be likely to double the degree of inhibition). Assuming, in the absence of data, that Lomentaria articulata has a selectivity factor similar to that of Porphyra ( Uemura et al. 1996 ) and that growth rate is proportional to photosynthetic carbon fixation, we might hypothesize a 7% decrease in growth rate under a doubling of [O2] and a 10% decrease in growth under a doubling of [O2] and a pre-industrial [CO2], relative to ambient conditions. Changes in growth rate of this magnitude have been detected in previous experiments with this and related species ( Kübler & Raven 1994; Kuebler, Davison & Yarish 1991).
Lomentaria articulata was collected from Fife Ness, East Neuk of Fife, UK, and vegetatively propagated in the laboratory at 10°C, 40–50 μmol photon m−2 s−1 (400–700 nm), under a 12 : 12 light : dark cycle for at least one week before being used in growth experiments. During this time, plants were grown in 3 dm3 jars of aerated Provasoli’s enriched seawater (PES, Provasoli 1968). The media were changed weekly.
For aeration of cultures with experimental gas mixtures, a factorial design of four concentrations each of CO2 and O2 (20, 90, 190 and 370 mmol O2 mol−1 and 235, 350, 700 and 1750 μmol CO2 mol−1), in N2 at approximately atmospheric pressure were supplied by an open flow gas mixing system built by R. Parsons and described in detail in Parsons, Raven & Sprent (1992). Gas concentrations in each outflow were monitored using a Clark-type O2 electrode and an infrared gas analyser. Each gas mixture was used to aerate two replicate tubes of 60 cm3 CO2-free PES and allowed to equilibrate before inoculation. Tubes were then inoculated with 100–200 mg Lomentaria articulata branches and incubated at 10 °C with constant aeration for 3 weeks. The small amount of tissue in each culture and the illumination of culture tubes from the side ensured that self-shading was minimal. The PFD during incubations, 40 μmol photon m−2 s−1, was saturating for growth ( Kübler & Raven 1994).
One half of the medium in each tube was replaced weekly with fresh CO2-depleted PES. DIC concentrations of all cultures were measured periodically by injecting a sample of medium into an acid chamber, inline with the CO2-free air intake of an infrared gas analyser. The pH of each culture was measured each time samples were taken. The equilibrium concentrations of DIC species were calculated according to Stumm & Morgan (1996). pH values measured in the incubation media were within 0·06 pH units of those predicted based on complete equilibrium of CO2 between the gaseous and liquid phases.
At the end of the incubation period, a sample of medium from each tube was fixed with 1 cm3 dm−3 saturated HgCl2 solution and sealed in an airtight vial for subsequent determination of the source stable carbon isotope ratio based on DIC. The biomass in each tube was blotted, weighed and dried overnight. Dry biomass was weighed and ground to a fine powder for chemical analysis.
Biomass-based relative growth rates were calculated from the natural logarithms of fresh biomass of the inoculum for each tube and the biomass at the end of 3 weeks. Growth rates on a carbon basis were calculated from the carbon content of the dry biomass of the inoculum and the biomass in culture at the end of the growth period.
Carbon and nitrogen content and stable carbon isotope ratios of dried organic matter at the end of the growth experiments were simultaneously determined on 1 mg dried samples using a combined CHN analyser (Carlo-Erba) and mass spectrometer (VG-SIRA). Source inorganic carbon samples were prepared by stripping the DIC from 5 cm3 of preserved culture medium with 100 mm3 phosphoric acid in evacuated 10 cm3 tubes with septum tops. The gas evolved was analysed using the mass spectrometer with helium as the carrier gas. Stable carbon isotope ratios of the source gas cylinders were measured directly using 10 cm3 samples at atmospheric pressure.
The entire experiment was run twice, once in the summer and once in the winter, and the results combined to give four replicate cultures for each gas mixture. The results for each character measured were analysed using two-way ANOVA with [CO2] and [O2] as the classification variables, with two replicates nested within each experimental date. Percentage data were arc-sin transformed prior to ANOVA. Non-significant sources of variation, identified in the ANOVA, were examined for post hoc pooling at α = 0·2 as described by Winer (1971). One-dimensional contrasts were used to test specifically for significant effects of increased and decreased [CO2] and [O2] relative to the current ambient values.
Lomentaria articulata had a positive growth rate, in terms of daily net carbon accumulation, under all combinations of [CO2] and [O2] tested ( Fig. 1). Relative carbon growth rate was affected by [CO2] (F3,25 = 6·87, P = 0·01) but not by [O2] (F3,25 = 0·24, P > 0·80), and there was no significant interactive effect of [CO2] and [O2] (F9,16 = 0·55, P > 0·81). The effect of [CO2] on growth rate was due to the effect of increased [CO2] (F1,25 = 4·36, P < 0·025) as opposed to decreased [CO2] (F1,25 = 0·96, P > 0·25) relative to the current ambient level (≈ 358 μmol CO2 mol−1), with relative growth enhancements of 52% and 23% under double and fivefold [CO2] enrichment, respectively. There was a significant difference in carbon growth rate between the two experimental dates (F16,32 = 2·84, P < 0·006).
Growth rate in terms of wet biomass ( Fig. 2) was significantly affected by [CO2] (F3,9 = 5·83, P < 0·02), particularly increasing [CO2] (F1,9 = 5·93. P < 0·05) rather than lowering [CO2] (F1,9 = 1·30, P > 0·27) relative to the ambient level and unaffected by [O2] (F3,9 = 0·13, P > 0·93), with no significant interaction between [CO2] and [O2] effects on growth rate (F9,16 = 0·333, P > 0·95). In contrast to the carbon growth rate, the growth rate based on wet biomass was low or negative at [CO2] below the growth optimum and there was no effect of experimental date (F16,32 = 1·42, P > 0·19). The mean growth enhancements due to changing [CO2], assuming no effect of [O2], were + 314% and + 50% at [CO2] of 2× and 5× ambient levels, respectively.
At the end of the growth period, all of the experimental material had lower wet to dry biomass ratios ( Fig. 3) than did the inoculum (mean 7·28, SE 0·61). This indicated that growth conditions in the laboratory increase dry matter content relative to field conditions as has been observed previously ( Kübler & Raven 1994, 1996). There was a significant effect of [CO2] (F3,25 = 3·71, P < 0·025) but not [O2] (F3,25 = 0·13, P > 0·93) on final wet to dry biomass ratios, such that cultures grown at lower [CO2] had relatively less dry matter than did those grown at high [CO2] ( Fig. 3). There was significant variation in dry matter content between the two experimental dates (F25,32 = 1·98, P < 0·03).
Most of the variation in the N content was due to differences between the starting material on the two experimental dates (F16,32 = 21·24, P < 0·001). There was a significant effect of [CO2] on %N (F3,9 = 25·62, P < 0·0001), no effect of [O2] (F3,9 = 0·87, P > 0·48) and no interactive effect of [CO2] and [O2] (F9,16 = 0·04, P > 0·99) on %N ( Fig. 4). Overall, %N at the end of the incubations was higher than in the inocula (mean 3·1% of dry weight, SE 0·1, cf. Fig. 4), which is characteristic of growth with saturating nutrient concentrations in the medium.
The carbon content of dry biomass was significantly affected by [CO2] (F3,9 = 4·68, P < 0·03) but not by [O2] (F3,25 = 0·36, P > 0·75) ( Fig. 5) such that the cultures grown at lower [CO2] had greater final carbon content per unit dry weight. In most treatments, mean C content at the end of the experiment was greater than that of the starting material (initial mean = 27·3% of dry weight, SE 0·21, cf. Fig. 5) and different between the two experimental dates (F16,32 = 5·87, P < 0·01).
The C and N contents responded to the different gas mixtures in a similar pattern. The slope of the C content response was greater than that of the N content response and sothe final C : N ratios were inversely related to [CO2] (F3,9 = 15·16, P < 0·0007), unaffected by [O2] (F3,9 = 1·25, P > 0·34) and differed between the two experimental dates (F25,32 = 11·42, P < 0·0001).
Stable carbon isotope discrimination
δ13C values of the organic matter (relative to the Pee Dee Belemnite (PDB)) at the end of the incubations were more negative under higher [CO2] (F3,25 = 19·16, P < 0·0001) and were unaffected by [O2] (F3,25 = 1·09, P > 0·35). There was a significant effect of experimental date on δ13C (F25,32 = 3·30, P < 0·0008) which was due to a difference in stable carbon isotope ratio of the compressed 5% CO2 gas supplied to the gas mixer.
Stable carbon isotope ratios were corrected for the source carbon isotope ratios of the DIC in the media according to the formula Δ = (δ13CDIC source–δ13Csample)/(1 + δ13Csample) to give values of discrimination (Δ13C in ‰ where all δ values were relative to PDB). Δ13C values increased with increasing [CO2] (F3,9 = 4·77, P < 0·04), indicating increasing discrimination when CO2 was readily available, and were unaffected by [O2] (F3,9 = 3·55, P > 0·06) ( Fig. 6). There were no significant interactions between the effects of [CO2] and [O2] (F9,16 = 0·37, P > 0·93). This indicated that the discrimination against 13C by Rubisco was unaffected by photorespiration under high [O2] even under reduced [CO2] ( Gillon & Griffiths 1997). In general, the values of Δ13C were lower than is typical of L. articulata grown at air equilibrium (Δ13C ≈ 30–33‰) ( Kübler & Raven 1994, 1996; Raven et al. 1995 ). This suggested that not all the organic carbon present in the inocula (fixed from a total DIC source at air equilibrium ≈ 0‰ relative to PDB) was turned over before the end of the experiment. Assuming that gas composition had no effect on the relative turnover rates of stored and newly synthesized organic carbon, this bias in Δ13C due to residual organic carbon from the inocula would be greater at low [CO2] where proportionately less carbon was fixed during the experimental period ( Fig. 1). However, the pattern of carbon isotope discrimination in response to changing gas composition did not mirror that of relative growth rate ( Fig. 6 cf. Figs 1 & 2) at the highest [CO2]. Therefore, the greater discrimination under greater [CO2] was not solely an effect of the growth rate diluting the organic carbon fixed before the experiment.
Implications for growth under elevated CO2
Atmospheric CO2 concentrations equivalent to the lowest [CO2] used here were not unusual during past interglacial periods ( Jouzel et al. 1993 ) and not much lower than those prevailing less than two centuries ago before the industrial revolution ( Barnola et al. 1987 ). While the limitations of extrapolating from laboratory growth data to the natural environment should not be underestimated, it is tempting to make some cautious observations. Lomentaria articulata had a growth rate under a pre-industrial [CO2] which was statistically indistinguishable from its growth rate under the current [CO2] suggesting that this species may not have benefited in terms of growth from recently elevated atmospheric [CO2], although it might be expected to do so eventually. Unfortunately, it is not possible to know whether L. articulata was as abundant in the pre-industrial North-east Atlantic intertidal zone as it is today.
Enhancements of algal growth rates with increasing [CO2] have been reported previously ( Gao et al. 1993 ; Negoro et al. 1991 , 1993). Two factors complicate comparisons between those studies and this one. Firstly, in earlier studies, the [CO2] used was significantly higher than would ever be expected atmospherically, either 50 mmol CO2 mol−1 air or power plant flue gas. Here, the relative growth enhancement under 5× current atmospheric [CO2] (< 2 mmole CO2 mol−1 air), was significantly less than that under 2× [CO2] and the response curve was markedly non-linear. Non-linearity of response to elevated [CO2] is commonly observed in terrestrial systems in cases where three or more levels of [CO2] are used ( Körner 1995). Secondly, the algal species used by Gao et al. (1993) and by Negoro et al. (1991 , 1993) are able to use bicarbonate as an inorganic carbon source for photosynthesis, unlike L. articulata. Although the dissolved inorganic carbon pool would be relatively enriched in CO2 relative to HCO3− at the low pH caused by aeration with 5% CO2, the maximum rate of inorganic carbon supply to Rubisco could differ widely between organisms with different mechanisms of inorganic carbon acquisition. To our knowledge, this is the first experimental study of seaweed responses to environmentally relevant changes in atmospheric gas composition (but see Beardall et al. 1998 for a theoretical treatment). In a study which also employed both elevated and reduced [CO2], the pattern of response of net photosynthesis and growth of a C3 terrestrial plant to changing atmospheric CO2 ( Dippery et al. 1995 ; Tissue et al. 1995 ) was roughly parallel to that of growth of L. articulata ( Fig. 1).
Inherently slow-growing species
In terrestrial plants, there is an inverse correlation between growth rate and relative growth enhancement due to elevated [CO2] ( Poorter 1993; Bowler & Press 1996), such that plants which have a high ratio of maintenance respiration to photosynthetic rates tend to have higher relative growth enhancements with elevated [CO2] ( Lloyd & Farquhar 1996). The pattern of low absolute growth rate but high relative enhancement in response to elevated external CO2 in the case of Lomentaria articulata is similar to that exhibited by such slow-growing terrestrial species. Lomentaria species tend to be slow-growing relative to many other intertidal seaweeds ( Kuebler et al. 1991 ; Kübler & Raven 1994, 1996). We have previously found that growth rates of L. articulata and other species of Lomentaria are generally low and change only slightly, in absolute terms, in response to changes in light level ( Kübler & Raven 1994), dynamics of light supply ( Kübler & Raven 1996) and temperature ( Kuebler et al. 1991 ). This generally refractory rather than environmentally plastic response pattern was described previously for ‘inherently’ slow-growing terrestrial plants ( Grime, Crick & Rincon 1986). The difference in responses of aquatic and terrestrial plants to changing [CO2] in the surrounding medium may be related to the greater variability of aquatic [CO2] due to the much slower diffusion of CO2 in water than in air.
Lomentaria species, being finely branched and hollow, do not have the morphological characteristics associated with previously described inherently slow-growing seaweeds, such as high frond-specific carbon content or a high ratio of non-photosynthetic to photosynthetic tissue ( Hanisak, Littler & Littler 1988; Markager & Sand-Jensen 1996). Lomentaria articulata does exhibit physiological characteristics of inherently slow-growing species such as decoupling of growth rate and carbon content under elevated [CO2] (this study) and at high light levels ( Kübler & Raven 1994). The relative growth rate of Lomentaria articulata under optimal conditions, which averaged only 5·8% day−1 in terms of biomass and 2·3% day−1 in terms of carbon, was not high, relative to many other seaweeds ( Markager & Sand-Jensen 1996) or to the congeneric species L. baileyana ( Kuebler et al. 1991 ). C content of L. articulata was inversely related to [CO2], the converse of the general pattern observed for C3 terrestrial plants ( Tissue et al. 1995 ). The fact that tissue C content was lower and growth rates were greater at elevated CO2 is not consistent with our original model of growth limitation by the simple diffusion of inorganic C from the media to the chloroplasts in the absence of a CCM. Although all the cultures accumulated tissue carbon during the experiment, the greatest growth limitation occurred in the cultures with large accumulations of tissue C. This supports the notion that there is some inherent limitation on the conversion of photosynthate to new tissue.
The general oxygen insensitivity of Lomentaria articulata is somewhat difficult to explain given the presumed mode of inorganic carbon acquisition and fixation in this species. Lomentaria articulata is among the relatively small number of freshwater and marine phototrophs for which gas exchange characteristics are consistent with a model of diffusion of CO2 to the site of Rubisco ( Raven & Beardall 1981; Maberly 1990; Johnston et al. 1992 ; Maberly et al. 1992 ; Raven et al. 1995 ). However, it is important to note that the ‘evidence’ for diffusive uptake of CO2 by the above marine phototrophs is actually a lack of evidence for modes of carbon acquisition other than CO2 diffusion, which, being the mechanistically most simple case, is assumed to be the evolutionary default. Here, the patterns of stable carbon isotope discrimination in response to changing [CO2] showed no evidence of discrimination related to photorespiration ( Gillon & Griffiths 1997) or of induction of CCMs ( Raven et al. 1995 ) at low [CO2] : [O2] ratio. Assuming that the Rubisco of L. articulata is similar to the highly conserved Rubisco of higher plants, cyanobacteria and other red algae ( Portis 1992; Raven 1997a) and diffusive entry of CO2 and O2 into the tissue, it was reasonable to expect a significant inhibition of net photosynthesis and potentially growth rate in response to a 10-fold range of [O2] in the aeration stream. Given the unlikelihood of a completely novel Rubisco in L. articulata and the wide variety of carbon-concentrating mechanisms employed by aquatic organisms ( Raven 1997a, b), the lack of an O2 effect suggested that conditions at the site of Rubisco must be different from that predicted by diffusion alone or that L. articulata has a low oxygen affinity Rubisco such as has been described for the diatom, Phaeodactylum tricornutum ( Badger et al. 1998 ).
Raven (1997b) showed that it was energetically possible for algal cells to maintain, through compartmentalization, an internal disequilibrium of dissolved inorganic C species such that the [CO2] around Rubisco is higher than that predicted by diffusive uptake of CO2 with no net intracellular accumulation of inorganic C. In the case of Lomentaria articulata, the responses of growth and carbon uptake characteristics including the stable carbon isotope ratios are consistent with a non-saturating and variable [CO2] in the chloroplasts under atmospheric [CO2] up to double the current ambient value. However, the mechanistic basis for such a modulation [CO2] is currently unknown.
The authors are indebted to Richard Parsons for the design and construction of the gas mixing system. Steve Dudgeon assisted with the statistical analyses. Steve Long and anonymous reviewers contributed to the improvement of the manuscript. This work was supported by NERC (UK).