Geobiology (2014)

DOI: 10.1111/gbi.12112

Isotopic discrimination and kinetic parameters of RubisCO from the marine bloom-forming diatom, Skeletonema costatum A. J. BOLLER,1 P. J. THOMAS,1 C. M. CAVANAUGH2 AND K. M. SCOTT1 1 2

Department of Integrative Biology, University of South Florida, Tampa, FL, USA Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA, USA

ABSTRACT The cosmopolitan, bloom-forming diatom, Skeletonema costatum, is a prominent primary producer in coastal oceans, fixing CO2 with ribulose 1,5-bisphosphate carboxylase/oxygenase (RubisCO) that is phylogenetically distinct from terrestrial plant RubisCO. RubisCOs are subdivided into groups based on sequence similarity of their large subunits (IA–ID, II, and III). ID is present in several major oceanic primary producers, including diatoms such as S. costatum, coccolithophores, and some dinoflagellates, and differs substantially in amino acid sequence from the well-studied IB enzymes present in most cyanobacteria and in green algae and plants. Despite this sequence divergence, and differences in isotopic discrimination apparent in other RubisCO enzymes, stable carbon isotope compositions of diatoms and other marine phytoplankton are generally interpreted assuming enzymatic isotopic discrimination similar to spinach RubisCO (IB). To interpret phytoplankton d13C values, S. costatum RubisCO was characterized via sequence analysis, and measurement of its KCO2 and Vmax, and degree of isotopic discrimination. The sequence of this enzyme placed it among other diatom ID RubisCOs. Michaelis-Menten parameters were similar to other ID enzymes (KCO2 = 48.9
2.8 lM; Vmax = 165.1
6.3 nmol min1 mg1). However, isotopic discrimination (e = [12k/13k  1] 9 1000) was low (18.5&; 17.0–19.9, 95% CI) when compared to IA and IB RubisCOs (22–29&), though not as low as ID from coccolithophore, Emiliania huxleyi (11.1&). Variability in e-values among RubisCOs from primary producers is likely reflected in d13C values of oceanic biomass. Currently, d13C variability is ascribed to physical or chemical factors (e.g. illumination, nutrient availability) and physiological responses to these factors (e.g. carbon-concentrating mechanisms). Estimating the importance of these factors from d13C measurements requires an accurate e-value, and a mass-balance model using the e-value for S. costatum RubisCO is presented. Clearly, appropriate e-values must be included in interpreting d13C values of environmental samples. Received 8 August 2014; accepted 30 August 2014 Corresponding author: K. M. Scott. Tel.: 813 974-5173; fax: 813 974-3263; e-mail: [email protected]

INTRODUCTION Diatoms are major primary producers in the ocean, responsible for 20% of global marine carbon fixation (Armbrust, 2009). In areas of nutrient upwelling, large blooms of diatoms quickly dominate the phytoplankton communities and form the base of bloom-associated food webs (Dore et al., 2008; Armbrust, 2009). The diatom, Skeletonema costatum, is a common, global member of the plankton community in temperate areas (Eilertsen & Degerlund, 2010; Li et al., 2010) and has been well-characterized (e.g. Gervais & Riebesell, 2001; Brutemark et al., 2009;

© 2014 John Wiley & Sons Ltd

Trimborn et al., 2009); thus, this species is an important candidate for characterization of diatom carbon fixation kinetics and stable isotope discrimination. To catalyze carbon fixation, diatoms use ribulose 1,5-bisphosphate carboxylase/oxygenase (RubisCO). Five forms of RubisCO catalyze the carboxylase step of the Calvin– Benson–Bassham cycle (IA, IB, IC, ID, and II; Fig. 1; Tabita et al., 2007). Form I RubisCOs consist of eight large and eight small subunits, while form II enzymes are active as dimers of a single subunit which is homologous to form I large subunits. Form ID is particularly prevalent in marine habitats, where it is present in almost all nongreen algae

1

2

A. J . BOLLER et al.

Fig. 1 Maximum likelihood analysis of selected RubisCO large subunit genes. Form ID genes are from diatoms unless marked otherwise. Form II RubisCO genes were used to form the out-group of the tree, and bootstrap values from 1000 resamplings are indicated at the nodes. The scale bar for branch length indicates number of substitutions per site. GenBank accession numbers are provided in Table S1.

© 2014 John Wiley & Sons Ltd

Isotope discrimination by diatom RubisCO such as diatoms, coccolithophores, rhodophytes, and some dinoflagellates. Form IB RubisCO is the most extensively studied, as it is present in eukaryotic green chloroplasts from algae and land plants. Form IA is also prevalent in marine habitats, found in marine Prochlorococcus and Synechococcus spp. and many proteobacteria, and form IC has been detected in marine samples using molecular methods. Form II RubisCO is present in peridinin-containing dinoflagellates as well as many marine chemolithoautotrophic bacteria (Tabita et al., 2007). Diverse environmental and physiological pressures on carbon fixation have selected for RubisCO enzymes with kinetic properties which vary greatly. For example, CO2 is the form of dissolved inorganic carbon fixed by RubisCO (Cooper & Filmer, 1969); KCO2 values of form I RubisCO range from 5 to 175 lM, while those of form II range from 100 to 250 lM, which reflects the high CO2 environments that some organisms carrying form II RubisCO inhabit (Horken & Tabita, 1999). These differences in kinetic properties also extend to the relative rates at which RubisCOs fix 12CO2 vs. 13CO2, which has the potential to exert a substantial influence on the stable carbon isotope composition of biomass (d13C = [(R/Rstd)  1] 9 1000, where R = 13C/12C, Rstd = 13C/12C of the Pee Dee Belemnite standard). The d13C of phytoplankton biomass has been used extensively to infer the physical and physiological factors influencing carbon fixation in the ocean (Freeman, 2001). d13C values of phytoplankton biomass are more 13C-depleted (d13C ~ 34& to 16&) than ocean dissolved carbon dioxide (d13C ~ 10 to 7&; Freeman, 2001; Mook et al., 1974), largely due to isotopic discrimination by RubisCO (e = RCO2/Rfixed  1) 9 1000), as all RubisCOs fix 12CO2 more rapidly than 13CO2 (Hayes, 2001). As the e-value of a particular RubisCO enzyme is constant, any process that changes the d13C of the CO2 pool available to RubisCO will also alter the d13C value of the carbon that is fixed, that is, d13Cfixed will track d13CCO2 with an interval approximately equal to e. Accordingly, the d13C

3

of environmental CO2 can affect phytoplankton biomass d13C values (Freeman, 2001). Furthermore, physiological factors such as CO2-concentrating mechanisms (CCMs) also have an impact, as they can have a substantial influence on the d13C value of intracellular CO2. While the total concentration of DIC in seawater is approximately 2 mM, due to DIC speciation relative to pH, only approximately 20 lM is in the form of CO2 (Millero & Sohn, 1992), which is smaller than the KCO2 of most RubisCO enzymes. Consequently, many diatoms use a CCM to supplement the supply of CO2 to RubisCO (Roberts et al., 2007). For some diatoms, both CO2 and HCO 3 uptake have been implicated, with carbonic anhydrase-mediated CO2 supply via dehydration of HCO 3 proposed to occur adjacent to RubisCO in the chloroplast (Hopkinson et al., 2011). CO2 and HCO 3 uptake systems are induced when the carbon demand by cell growth exceeds the rate of CO2 supply by diffusion and are likely responsible for the shifts in d13C observed under different growth rates (Trimborn et al., 2009). Further, a few species may also have a C4-like pathway, which could also impact biomass d13C values (Roberts et al., 2007). To estimate the magnitude of these factors on biomass d13C values of autotrophs, it is common to assume the RubisCO e-value is similar to that measured for spinach, approximately 25–29& (e.g. Bidigare et al., 1997; Popp et al., 1998; Burkhardt et al., 1999). This assumption is problematic, as the few e-values that have been measured for RubisCOs from different organisms vary considerably. Form IA and IB RubisCO enzymes studied to date fractionate more (e = 22–30&; Table 1; Guy et al., 1993; Scott et al., 2004b, 2007) than form II (e = 18–19.5&; Roeske & O’leary, 1985; Robinson et al., 2003). Discrimination by form ID, measured to date for a single organism, the coccolithophore Emiliania huxleyi, was extremely low (e = 11&; Boller et al., 2011). As organisms with form ID RubisCO are ubiquitous and abundant in the ocean, the lower isotopic discrimination by form ID RubisCO could greatly influence

Table 1 RubisCO e-values measured with high-precision methods Form of RubisCO

Taxon

Organism type

e-value (&)

References

IA IB

Solemya velum symbiont Prochlorococcus marinus MIT9313* Spinacia oleracea

c-Proteobacterium Cyanobacterium Plant

Cyanobacterium Primarily Proteobacteria Coccolithophore Diatom c-Proteobacterium a-Proteobacterium

Scott et al. (2004b) Scott et al. (2007) Roeske & O’leary (1985) Guy et al. (1993) Scott et al. (2004b) Guy et al. (1993)

IC

Anacystis nidulans Many

24.5 24 29–30 30 28 22 ND 11.1 18.6 19.5 18

Boller et al. (2011) This study** Robinson et al. (2003) Guy et al. (1993)

ID II

Emiliania huxleyi* Skeletonema costatum* Riftia pachyptila symbiont Rhodospirillum rubrum

*Marine phytoplankton. **These data are the result from this study and are boldfaced accordingly.

© 2014 John Wiley & Sons Ltd

4

A. J . BOLLER et al.

13

C content in marine organic carbon and may even account for the observed 13C enrichment of marine carbon. The RubisCO enzyme of S. costatum was characterized to provide a better baseline upon which to decipher the influences of environmental and physiological factors on diatom d13C values. This will inform efforts to elucidate autotroph physiology, marine trophic interactions, and to model ocean carbon fixation values based on stable isotopes (Post, 2002; Loubere & Bennett, 2008; Rayner et al., 2008). Further, the RubisCO e-value for a given species is needed to evaluate d13C values of organisms employing a CCM. To this end, a cell-level carbon flux model was developed, which uses the e-value determined for S. costatum RubisCO to predict biomass d13C value deviations caused by induction of a CCM.

MATERIALS AND METHODS Diatom cultivation Skeletonema costatum cultures (CCMP1332; ProvasoliGuillard National Center for Culture of Marine Phytoplankton) were grown in f/2 medium at 18–22 °C, under constant illumination (Guillard & Ryther, 1962). One litre of f/2 media was inoculated with 5 mL of cell culture. After 1 week of growth, the 1 L culture was added to 4 L of fresh media for another week of growth with aeration. Cells were harvested from each 5 L culture by centrifugation (50009 g, 15 min), flash-frozen separately in liquid nitrogen, and stored at 80 °C. RubisCO large subunit gene (rbcL) sequence analysis To verify and extend available sequences (GenBank JN159931 and FJ002107), the rbcL gene of S. costatum was resequenced. DNA was purified from S. costatum using the CTAB method (http://my.jgi.doe.gov/general/proto cols/DNA_Isolation_Bacterial_CTAB_Protocol.doc). Primers for S. costatum rbcL were designed from a partial sequence (Daugbjerg & Andersen, 1997) and purchased from Integrated DNA Technologies (Coralville, IA, USA). To amplify the first half of the gene (1–840 bp), the forward primer 50 -GGGTTACTGGGATGCTTCATACAC-30 and reverse primer 50 -CCAACAGCTTTAGCATACTCAGCAC30 were used. The primers used to amplify the other half of the gene (560–1428 bp) were forward primer 50 -GGAA GGTATTAACCGTGCATCAGC-30 and reverse primer 50 -TCTGTTTGCAGTTGGTGTTTCAGC-30 . Amplicons were ligated into pCR2.1, propagated in Top10 cells (Invitrogen Inc., Carlsbad, CA, USA), and sequenced using standard molecular techniques. The sequence was deposited in GenBank (accession # KM594531. Sequences encoding RubisCO large subunits (rbcL or cbbL, depending on organism) were obtained from NCBI

via BLASTp and aligned via MUSCLE (Edgar, 2004). The alignment was manually verified with respect to key catalytic residues (Tabita et al., 2007) and refined via gBLOCKS using stringent criteria (Talavera & Castresana, 2007). A maximum likelihood tree was generated from this alignment using MEGA 4.0 with 1000 replicates for bootstrapping (Tamura et al., 2011). RubisCO purification from S. costatum RubisCO was partially purified from frozen cell pellets using ammonium sulfate precipitation. For all purification steps, samples were kept at 4 °C. Frozen cell pellets were resuspended in lysis buffer [20 mM Tris pH 7.5, 10 mM MgCl2, 5 mM NaHCO3, 1 mM EDTA, and 1 mM dithiothreitol (DTT)], sonicated twice for 30 s with glass beads, and centrifuged (5000 9 g, 15 min, 4 °C). (NH4)2SO4 was added to the supernatant to a final concentration of 30% saturation, and proteins were allowed to precipitate for 15 min. After centrifugation (5000 9 g, 15 min, 4 °C), the pellet was discarded and the supernatant was brought to 50% (NH4)2SO4 saturation. Proteins that precipitated at 30–50% (NH4)2SO4 saturation were collected by centrifugation and dissolved in BBMD buffer (50 mM bicine pH 7.5, 25 mM MgCl2, 5 mM NaHCO3, and 1 mM DTT). The RubisCO extracts were desalted into BBMD buffer (HiPrep 26/10 desalting column; Amersham Biosciences, Piscataway, NJ, USA) and used immediately. RubisCO activity in extracts was monitored by the incorporation of 14CO2 into 3-phosphoglycerate (PGA). 0.5 mL of these RubisCO extracts was added to 1 mL of assay buffer (50 mM bicine, pH 8.0, 10 mM MgCl2, 1 mM EDTA, 5 mM DTT, and 2 lCi mL1 DI14C; MP Biomedicals, Irvine, CA, USA) and pre-incubated for 15 min. Reactions were started by the addition of 1 mM ribulose 1,5-bisphosphate (RuBP; Sigma R0878, St. Louis, MO, USA). At 30 s intervals, 350 lL portions were removed from the reaction mix and added to 200 lL glacial acetic acid to stop the reaction and drive off unincorporated DI14C. After the samples were dried by sparging with air, 3 mL of ScintiSafe Plus 50% scintillation cocktail (Fisher Scientific #SX25-5, Hampton, NH, USA) was added. Assays lacking RuBP were performed as a control for CO2 fixation independent from RubisCO activity. Michaelis-Menten parameters The KCO2 and Vmax of S. costatum RubisCO were estimated from carbon fixation rates measured radiometrically (Scott et al., 2007). The assay buffer (50 mM bicine, pH 7.5, 30 mM MgCl2, 1 mM DTT) was prepared with trace CO2 and O2 concentrations as in (Scott et al., 2007) and sealed in glass vials with stir bars under an N2 overpressure with gastight septa. Buffer was supplemented with DIC to eight

© 2014 John Wiley & Sons Ltd

Isotope discrimination by diatom RubisCO concentrations ranging from 0.2 to 14 mM, and 2 lCi mL1 DI14C. Bovine erythrocyte carbonic anhydrase (40 lg mL1; Sigma C3934) was added to each vial to maintain DIC at chemical and isotopic equilibrium. Approximately 1 mg mL1 of S. costatum RubisCO extract in BBMD buffer was added (final concentration ~ 0.02 mg mL1), followed quickly by RuBP (0.4 mM final concentration) to begin the reaction. Samples (200 lL) were removed with a gastight syringe at 1 min intervals over a 4 min time course and immediately injected into scintillation vials containing 200 lL glacial acetic acid to quench the reaction and remove unfixed DIC. The samples were sparged with air for approximately 6 h to ensure that they were completely dry before 3 mL scintillation cocktail were added to quantify the acid-stable (fixed) 14C. The initial activity of each incubation was measured by injecting 10 lL samples into scintillation cocktail containing phenylethylamine (SX10-1000; Fisher Scientific) to trap the NaH14CO3. Both acid-stable 14C samples and the initial activities were measured via scintillation counting. Five independent experiments were conducted, with RubisCO extract from separate cultures. KCO2 and Vmax values for the five experiments were estimated from the carbon fixation rates using direct linear plots (Eisenthal & Cornish-Bowden, 1974), which treat each data point equally, unlike linearizations such as Lineweaver-Burk, which overemphasize data points from the lowest and highest substrate concentrations. RubisCO isotope discrimination The e-value of S. costatum RubisCO extract was measured at pH 7.5 using the high-precision substrate depletion method (Scott et al., 2004b; Boller et al., 2011). The reaction was prepared by sparging BBMD buffer with N2 to minimize the concentration of O2, and 1 mg 25 mL1 of bovine erythrocyte carbonic anhydrase (CA; Sigma 3934) was added to maintain DIC at chemical and isotopic equilibrium. RubisCO extract was added to the BBMD buffer to a final concentration of approximately 1 mg mL1 protein. This reaction mixture was then filter-sterilized (0.45 lm) and loaded into a heat-sterilized, septum-sealed 25 mL glass gastight syringe with a stir bar. After activating the RubisCO by incubating the enzyme for 15 min at 25 °C in the gastight syringe, filter-sterilized (0.22 lm) enzymatically synthesized RuBP (approximately 100 lmoles; Scott et al., 2004b) was injected to begin the reaction. The concentration and d13C value of DIC were measured as the RubisCO reaction consumed CO2. The [DIC] of triplicate samples for each timepoint was measured via gas chromatography (HP/Agilent 5890A, Santa Clara, CA, USA; SD < 0.1 mM; Scott et al., 2007). Samples were also acidified in gastight syringes with 43% phosphoric acid (1:4 ratio) to terminate the reaction and convert the DIC to CO2. Using a vacuum line, the DIC

© 2014 John Wiley & Sons Ltd

5

was cryodistilled from the samples for d13C determination using a gas inlet mass spectrometer at the Boston University Stable Isotope Laboratory (SD < 0.1&). Five independent e measurements were performed using the S. costatum RubisCO extracts. e-values were calculated by regressing the natural log of the concentration of DIC against the natural log of its isotope ratio (R = 13DIC/12DIC): lnR ¼ ððaCÞ1  1Þ  ln½DIC þ lnðRDIC0  ð½DIC0 ðð1=aCÞ1ÞÞ

1

Þ

where a is the kinetic isotope effect (12k/13k), and C = Rbicarbonate 9 RCO21, which accounts for the equilibrium isotope effect between CO2 and bicarbonate (Mook et al., 1974; Scott et al., 2004a). Values of e were then calculated from a-values, as e = (a  1) 9 1000. The average e-value from replicate experiments was calculated using a Pitman estimator (Scott et al., 2004a). In previous work with RubisCO from E. huxleyi, this technique for determining e-value accuracy was confirmed by running spinach RubisCO (Boller et al., 2011), as its e-value had been previously determined with high-precision methods (e = 28–30&; Roeske & O’leary, 1985; Guy et al., 1993). The value determined using our techniques (28.3&) was in agreement with these, indicating good analytical accuracy. The experiments with enzyme from E. huxleyi (Boller et al., 2011) and S. costatum (reported here) were conducted shortly after running spinach RubisCO. To determine whether alternate carboxylases were fixing CO2, incubations were also run without RuBP, as well as in the presence of 5 mM 3-phosphoglycerate, the product of the RubisCO reaction and the only organic compound anticipated to accumulate under these reaction conditions (Sigma P8877).

RESULTS Resequencing of the rbcL gene of Skeletonema costatum resulted in an improved sequence, adding 72 nucleotide residues to previously deposited sequences. Maximum likelihood analysis confirmed the placement of the S. costatum rbcL gene with other form ID RubisCO rbcL genes, including other diatom species, coccolithophores, and rhodophytes (Fig. 1). RubisCO was the only carboxylase activity detected in the S. costatum RubisCO extracts. In incubations to which these extracts were added, [DIC] decreased over time, and 14 C-DIC was fixed only in the presence RuBP, indicating carbon fixation by other carboxylases was not occurring (Fig. 2). Adding the product of the RubisCO reaction, 3-PGA, did not stimulate DIC consumption by any C4-carboxylases potentially present in the extracts (Fig. 3). The

6

A. J . BOLLER et al. –4.466

Carbon fixed (nmol)

10 + RuBP – RuBP

8

lnR

6

–4.481

2

–4.486 –4.491

0

2

4

6

8

10

12

0.8

0.9

1

1.1

1.2

1.3

1.4

1.5

ln [DIC] (mM)

Time (min) Fig. 2 Radiometric assay of Skeletonema costatum RubisCO extracts. Assays were conducted in the presence and absence of ribulose 1,5-bisphosphate (RuBP) to confirm that detectable carboxylase activity was due to RubisCO. 6 5

[DIC] (mM)

–4.476

4

0

4

Fig. 4 Isotopic fractionation of dissolved inorganic carbon (DIC) as CO2 is consumed by RubisCO from Skeletonema costatum and spinach. R is the isotope ratio of DIC in the incubation, and [DIC] is its concentration. The lines (dotted for S. costatum, solid for spinach) were fitted from the data using the Pitman estimator average e-values and are a linear regression of lnR on ln[DIC], as described in the methods. Spinach data are replotted from Boller et al. (2011).

global carbon cycle models, and understanding the isotope geochemistry of marine primary productivity.

3 2

+ RuBP – RuBP – RuBP + PGA

1 0

Open symbols: S. costatum Closedsymbols: spinach

–4.471

0

2

4

RubisCO gene sequence analysis and Michaelis-Menten parameters 6

8

10

Time (h) Fig. 3 The consumption of DIC over time by Skeletonema costatum RubisCO. Five independent incubations which included ribulose 1,5-bisphosphate (+ RuBP) were used to calculate the e-value of S. costatum RubisCO. The experiments conducted in absence of RuBP (RuBP) or with added 3phosphoglycerate (RuBP + PGA) lack measureable carboxylase activity.

Michaelis-Menten kinetic parameters were very consistent in three independent experiments (KCO2 = 48.9
2.8 lM; Vmax = 165
6 nmol min1 mg1, respectively). e-values (=([12k/13k]  1) 9 1000) for five independent S. costatum RubisCO extracts were determined. Each extract was prepared from a separate 100 L culture, and an e-value was determined for each extract (15.9&, 18.6&, 19.0&, 19.7&, and 20.2&; Fig. 4). To combine the five values to obtain the least biased average value, a Pitman estimator was used (e = 18.5&; 95% CI: 17.0–19.9&; Scott et al., 2004a).

DISCUSSION The work here is a major boon in the understanding of a globally relevant bloom-forming diatom. Resequencing of the S. costatum rbcL resulted in an improved sequence and better resolution of the phylogenetic relationship of the diatom form ID RubisCOs. The Michaelis-Menten parameters permit a much more informed discussion of the role of a CCM in this species. The e-value provides a springboard for deciphering RubisCO activity in situ, refining

The amino acid sequence of the S. costatum rbcL gene product is very similar to Thalassiosira pseudonana (96% identical), which is consistent with the membership of both diatom species within the Thalassiosirales, with more distantly related members of the Bacillarophyceae (e.g. Cylindrotheca sp., Phaeodactylum tricornutum) still having substantial sequence identity (up to 93%). These results are concordant with recent multilocus phylogenetic analyses of diatoms (Sorhannus & Fox, 2012). Beyond diatoms, other form ID enzymes sampled for this phylogenetic analysis from the rhodophytes and coccolithophorids are more divergent (approximately 82% sequence identity to S. costatum). The KCO2 of S. costatum RubisCO (48.9 lM) is similar to those collected from Cylindrotheca spp. (30 and 35 lM; Read & Tabita, 1994). The KCO2 of form ID RubisCO from the coccolithophore E. huxleyi is higher (72 lM; Boller et al., 2011), while those from rhodophytes can be considerably lower (5–20 lM; Read & Tabita, 1994). Given that S. costatum RubisCO KCO2 is over twice the concentration of CO2 in seawater (20 lM), it is not surprising that this organism employs a CCM, increasing the activities of bicarbonate uptake and extracellular carbonic anhydrase when CO2 concentrations are low (Trimborn et al., 2009). RubisCO isotope discrimination No other diatom RubisCO e-values had been determined with high-precision methods prior to this study. The value measured here for S. costatum (e = 18.5&) is far lower

© 2014 John Wiley & Sons Ltd

Isotope discrimination by diatom RubisCO than the value commonly assumed for enzymes from these organisms based on spinach (e = ~25–29&), though not as low as that of E. huxleyi (e = 11.1&, Table 1; Boller et al., 2011). Form ID RubisCO e-values for other marine photoautotrophs are likely to be heterogeneous, even beyond the 7& span between these two enzymes. Form ID RubisCO sequences from rhodophytes do not fall in a discrete clade (Fig. 1), which may indicate divergent kinetic parameters. Some rhodophyte enzymes may have higher e-values, even as high as form IB from terrestrial plants, as biomass collected from rhodophytes can have very negative d13C values, lower than 30& (Maberly et al., 1992; Raven et al., 2002). Furthermore, the RubisCOs of two rhodophytes (Galdieria partita and Cyanidium caldarium; Horken & Tabita, 1999) have very high specificity for CO2 as a substrate relative to O2. This is consistent with a higher e-value, as RubisCO substrate specificity may be correlated with isotopic selectivity (Tcherkez et al., 2006). RubisCO e-value and biomass d13C The e-value determined for S. costatum RubisCO provides a baseline to interpret isotope discrimination measured in cultures of this organism. Isotope discrimination by intact cells is calculated as ep(=[RCO2/Rbiomass  1] 9 1000). Values for intact cells were low (ep = 6.2–14&) when S. costatum was grown under low-CO2 conditions or stimulated with increased light intensity to increase growth rate (Hinga et al., 1994; Burkhardt et al., 1999). ep-values lower than enzyme e-values can be attributed to 13C-enriched intracellular CO2 pools. As growth rate increases, the rate of CO2 fixation by RubisCO approaches the rate of exchange between intracellular and extracellular CO2 pools (Scott, 2003). Because RubisCO fixes 12CO2 more rapidly than 13 CO2, a pool of isotopically enriched intracellular DIC is formed. Hence, isotopically enriched biomass is produced by RubisCO as it is drawing from a more isotopically enriched intracellular CO2 pool. Additionally, in low [CO2] environments, the S. costatum CCM will be induced, which diminishes CO2 loss from the cytosol and exacerbates the 13CO2enriching effect of RubisCO activity on intracellular CO2. Therefore, culture biomass values (ep = 6.2–14&) lower than enzyme values (e = 18.5&) can be attributed to lower rates of exchange between intracellular and extracellular CO2 pools, relative to carbon fixation rates. When such rates of exchange are high, due to elevated [CO2], or decreasing growth rate due to nutrient or light limitation, values for cells were higher (ep = 10.3–19 &; Hinga et al., 1994; Gervais & Riebesell, 2001), as under these conditions RubisCO’s impact on the isotopic composition of intracellular CO2 is mitigated by exchange of CO2 with the environment. This effect is also apparent in values (ep = 10.6–19.5&) measured in blooms where S. costatum, or other diatoms,

© 2014 John Wiley & Sons Ltd

7

were the primary phytoplankton species (e.g. Dehairs et al., 1997; Kukert & Riebesell, 1998). Similar to cultures, some variability in ep-values for diatom blooms can be attributed to changes in [CO2] or growth rates caused by environmental factors, such as temperature, nutrient limitation, photon flux density, or pH (Johnston, 1996; Riebesell et al., 2000; Brutemark et al., 2009). Consistent with these potential variables, bloom organisms tend to have lower ep (10.6–12.2&) at the end of the bloom, when CO2 demand is high due to elevated growth rates (Kukert & Riebesell, 1998). Cell carbon flux model Analysis of biomass d13C values informed by the RubisCO e value can be used to determine the degree to which the CCM is induced in situ, to understand CCM relevance to bloom dynamics, and potential competitive advantage. With the measured e-value of S. costatum RubisCO, the d13C values of intracellular CO2 can be estimated, and with them, the relative rates of CO2 exchange and fixation, and the relative importance of CO2-concentrating mechanisms in the environment or in cultures. CCMs and other physiological adaptations to enhance CO2 uptake have a profound influence on how phytoplankton interact with the environment. For example, CCMs affect transition metal consumption by cells (e.g. Zn+2 and Cd+2 for carbonic anhydrase activity; Park et al., 2007) and also make cells more sensitive to limitation by light due to the added energetic expense of cell structures (e.g. pyrenoids) and active DIC uptake (DeLaRocha, 2008). Accordingly, estimating the activities of CCMs is important for illuminating the magnitude of the role of diatoms in the geochemical cycles of transition metals and suggests their sensitivity to physical factors (illumination). The RubisCO e-value permits this estimation from biomass d13C values with a model of DIC influx and consumption by the cell. The current model for the CCM of S. costatum, necessary for modeling stable carbon isotope ratios in this organism, indicates a role for both CO2 supply via diffusion and HCO 3 supply via active transport. Diatom cell and chloroplast membranes are extremely permeable to CO2; therefore, CO2 diffusion (Hopkinson et al., 2011b) and HCO 3 uptake (Burkhardt, 2005) deliver inorganic carbon to the cytoplasm. Cytoplasmic carbonic anhydrase converts some of the CO2 to HCO 3 ; cytoplasmic CO2 and HCO enter the chloroplast via diffusion and active 3 transport, respectively (Hopkinson et al., 2011b). Within the chloroplast, starch-jacketed pyrenoids facilitate carbon fixation by the RubisCO with which they are packed, likely facilitated by carbonic anhydrase-mediated dehydration of pyrenoid HCO 3 to CO2 (Pankratz, 1964; Gao, 2004; Hopkinson et al., 2011b; Sinetova et al., 2012). When CO2 is abundant enough to meet cellular demand, cell

8

A. J . BOLLER et al.

membrane and chloroplast CO2 uptake by diffusion dominate (Trimborn et al., 2009). Under low-CO2 or other conditions in which CO2 demand exceeds supply, S. costatum induces its CCM, demonstrating an increased reliance on HCO 3 and lower whole-cell KCO2 values (Burkhardt, 2005; Trimborn et al., 2009). A mass-balance-based model incorporating stable isotope values is presented to evaluate the degree to which the CCM is induced in this species, using d13C values from environmental or culture samples of S. costatum biomass, and the e-value of its RubisCO. The model depicted in Fig. 5 assumes that rapid exchange of DIC between the chloroplast and cytoplasm due to high chloroplast permeability to CO2 (Hopkinson et al., 2011b) prevents large differences in the d13C of the DIC in these two compartments from accumulating. Assuming steady-state conditions,

qfix þ qoCO2 þ qCACO2 ¼ qresp þ qiCO2 þ qCAbic

ð1Þ

qibic þ qCACO2 ¼ qCAbic

ð2Þ

which describe the interactions of carbon fixation rates (qfix), diffusion-mediated CO2 efflux (qoCO2) and influx (qiCO2), carbonic anhydrase activity (hydration = qCACO2; dehydration = qCAbic), respiration rates (qresp), and active bicarbonate uptake (qibic) on the intracellular CO2 (1) and HCO 3 (2) pools. Combining (1) and (2) results in qfix þ qoCO2 ¼ qresp þ qiCO2 þ qibic

ð3Þ

A mass-balance equation which combines these fluxes with the isotopic compositions of the pools of carbon upon which they are acting is

–

–

Fig. 5 Model of DIC uptake and fixation by Skeletonema costatum. Fluxes depicted include HCO 3 uptake (qibic), CO2 uptake (qiCO2), CO2 efflux (qoCO2), carbonic anhydrase-mediated dehydration (qCAbic), carbonic anhydrase-mediated hydration (qCA co2), CO2 fixation (qfix), and respiration (qresp). Isotope fractionation factors relevant to the fluxes are those for transport (eT), diffusion (eD), fixation by RubisCO (e), and the equilibrium isotope effect between dissolved CO2 and HCO 3 (ebic).

qfix ðdiCO2  eÞ þ qoCO2 ðdiCO2  eD Þ ¼ qresp ðdbiomass Þ þ qiCO2 ðdoCO2  eD Þ

ð4Þ

þ qibic ðdobic  eT Þ which accounts for the d13C value of intracellular (diCO2) and extracellular (doCO2) CO2, biomass (dbiomass), and extracellular HCO 3 (dobic), as well as the fractionation factor of RubisCO carboxylation (e), diffusion (eD), and transport (eT). Assuming that carbon fixation is the major source of organic carbon in biomass (diCO2  e = dbiomass) and that eT is likely to be small as it does not involve covalent bond formation (eT = eD), and expressing qoCO2 as a function of qfix, qresp, qiCO2, and qibic by rearranging equation (3), one obtains qfix  qresp ¼ ½qiCO2 ðdoCO2  dbiomass  eÞ þ qibic ðdobic  dbiomass  eÞ=ðeD  eÞ

ð5Þ

Dividing (5) by qiΣ (= qiCO2 + qibic), ðqfix  qresp Þ=qiR ¼ ½epCO2  e þ ðfbic  ebc Þ=ðeD  eÞ

ð6Þ

in which epCO2 = doCO2  dbiomass, fbic = qibic/qiΣ, and ebc is the equilibrium isotope effect between CO2 and HCO 3 (~d13CHCO3  d13CCO2). The relationship between (qfix  qresp)/qiΣ and epCO2values makes it possible to assess fbic, the fraction of DIC supply to the cell that is derived from HCO 3 uptake, which should be proportional to CCM activity. As qfix  qresp is approximately equal to the growth rate, the ratio (qfix  qresp)/qiΣ is the fraction of DIC that is assimilated into biomass (the rest is lost from the cell as diffusion). When the value of e determined here for S. costatum RubisCO is inserted into equation (6) and plotted for all possible fbic-values (Fig. 6A), using the full range of epCO2values measured (6–20&; Hinga et al., 1994; Burkhardt et al., 1999; Gervais & Riebesell, 2001), and ebc from Mook et al. (1974), it illustrates that the larger epCO2-values are consistent with DIC supply rates (qiΣ) outpacing growth (qfix  qresp) and more reliance on CO2 (smaller fbic values). Smaller epCO2-values are consistent with growth rates approaching DIC supply rates and more reliance on HCO 3 ; in turn, these smaller epCO2-values, suggesting a greater reliance on a CCM, would also suggest a higher demand for the transition metals and illumination necessary for CCM function. The importance of using the correct e-value to interpret autotroph d13C values is well illustrated by comparing the above results, using the measured e-value for S. costatum RubisCO (18.5&), to results using the commonly used value from spinach RubisCO (29&; Fig. 6B; Roeske & O’leary, 1984). Using the spinach RubisCO e-value significantly inflates the range of values of fbic at each epCO2 and

© 2014 John Wiley & Sons Ltd

Isotope discrimination by diatom RubisCO A

B

9

dominant phytoplankton species during bloom development, isotopically less selective RubisCO can lead to d13C enrichment throughout marine food webs. For example, zooplankton from areas with diatom blooms have relatively 13 C-enriched biomass compared to zooplankton found near phytoplankton with 13C-depleted biomass (Fry & Wainright, 1991). If RubisCO enzymes from other marine phytoplankton also have lower e-values, as in S. costatum and E. huxleyi (Table 1), this could help to explain why marine organic carbon is 13C-enriched relative to terrestrial C3-based ecosystems. The few e-values that have been collected demonstrate that isotopic discrimination by RubisCO varies substantially among organisms. Each e-value can individually be used to interpret d13C values of biomass from its host organism. At this point, it is unwise to apply a single e-value in a blanket manner across taxa, and even within a particular RubisCO form. Instead, if the dominant organism(s) in a sample have not had their RubisCO e-values determined, it would be better to interpret biomass d13C values using the full range of RubisCO e-values measured thus far.

ACKNOWLEDGMENTS

Fig. 6 CO2-concentrating mechanism activity and carbon demand to supply ratio ([qfix  qresp]/qiΣ) predicted from ep (=d13CCO2  d13Cbiomass) and equation (6). fbic = fraction of inorganic carbon supply to the cell as bicarbonate. Each line between those demarcating fbic = 0 and fbic = 1 represents an increment in fbic of 0.1. (A) Values calculated using e = 18.5& (Skeletonema costatum RubisCO). (B) Values calculated assuming the spinach e (29&).

We are grateful to NSF Biological Oceanography for their support (BIO-OCE 0327488 to C.M.C. and K.M.S.), and to the anonymous reviewers, whose suggestions improved the manuscript. Special thanks are given to undergraduate researchers Kelly Fitzpatrick and Michelle Echevarria for their assistance with the experiments.

REFERENCES therefore overestimates the magnitude of the contribution of the CCM to DIC supply for carbon fixation. Overestimating the activity of the CCM could result in an overestimation of the impact of CCM-relevant parameters (e.g. irradiance, transition metal availability) on in situ carbon fixation rates, and overestimating the effect of increases in atmospheric CO2 on competition among phytoplankton. RubisCO e-values and food webs Ascribing changes in bloom biomass d13C values to variations in environmental CO2 concentrations and the carbon-concentrating mechanisms induced in response to them may sometimes be misguided. The heterogeneity in the e-values that have been measured for three phytoplankton carboxylases (Table 1) indicates that bloom d13C values will also be influenced by the taxonomic composition of the bloom phytoplankton. Changes in biomass d13C resulting from displacement by phytoplankton with different RubisCO e-values would ripple through trophic levels. As S. costatum and other diatoms can quickly become the

© 2014 John Wiley & Sons Ltd

Armbrust EV (2009) The life of diatoms in the world’s oceans. Nature 459, 185–192. Bidigare RR, Fluegge A, Freeman KH, Hanson KL, Hayes JM, Hollander D, Jasper JP, King LL, Laws EA, Milder J, Millero FJ, Pancost R, Popp BN, Steinberg PA, Wakeham SG (1997) Consistent fractionation of 13C in nature and the laboratory: growth rate effects in some haptophyte algae. Global Biogeochemical Cycles 11, 279–292. Boller AJ, Thomas PJ, Cavanaugh CM, Scott KM (2011) Low stable carbon isotope fractionation by coccolithophore RubisCO. Geochimica et Cosmochimica Acta 75, 7200–7207. Brutemark A, Lindehoff E, Gran
eli E, Gran
eli W (2009) Carbon isotope signature variability among cultured microalgae: influence of species, nutrients and growth. Journal of Experimental Marine Biology and Ecology 372, 98–105. Burkhardt S, Riebesell U, Zondervan I (1999) Stable carbon isotope fractionation by marine phytoplankton in response to daylength, growth rate, and CO2 availability. Marine EcologyProgress Series 184, 31–41. Chen XW, Gao KS (2004) Roles of carbonic anhydrase in photosynthesis of Skeletonema costatum. Journal of Plant Physiology and Molecular Biology 30, 511–516. Cooper TG, Filmer D (1969) The active species of “CO2” utilized by ribulose diphosphate carboxylase. Journal of Biological Chemistry 244, 1081–1083.

10

A. J . BOLLER et al.

Daugbjerg N, Andersen RA (1997) A molecular phylogeny of the heterokont algae based on analyses of chloroplast-encoded rbcL sequence data. Journal of Phycology 33, 1031–1041. Dehairs F, Kopczynska E, Nielsen P, Lancelot C, Bakker DCE, Koeve W, Goeyens L (1997) d13C of Southern Ocean suspended organic matter during spring and early summer: regional and temporal variability. Deep sea Research. Part 2, Topical Studies in Oceanography 44, 129–142. Dore JE, Letelier RM, Church MJ, Lukas R, Karl DM (2008) Summer phytoplankton blooms in the oligotrophic North Pacific Subtropical Gyre: historical perspective and recent observations. Progress in Oceanography 76, 2–38. Drum RW, Pankratz HS (1964) Pyrenoids, raphes, and other fine structure in diatoms. American Journal of Botany 51, 405–418. Edgar RC (2004) MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Research 32, 1792–1797. Eilertsen HC, Degerlund M (2010) Phytoplankton and light during the northern high-latitude winter. Journal of Plankton Research 32, 899–912. Eisenthal R, Cornish-Bowden A (1974) The direct linear plot. Biochemistry Journal 139, 715–720. Freeman KH (2001) Isotopic biogeochemistry of marine organic carbon. In: Stable Isotope Geochemistry (eds Valley JW, Cole DR). Mineral. Soc. Amer., Washington, DC, pp. 579–605. Fry B, Wainright SC (1991) Diatom sources of carbon-13 rich carbon in marine food webs. Marine Ecology-Progress Series 76, 149–157. Gervais F, Riebesell U (2001) Effect of phosphorus limitation on elemental composition and stable carbon isotope fractionation in a marine diatom growing under different CO2 concentrations. Limnology and Oceanography 46, 497–504. Guillard RRL, Ryther JH (1962) Sudies of marine planktonic diatoms. I. Cyclotella nana Hustedt and Detonula confervacea Cleve. Canadian Journal of Microbiology 8, 229–239. Guy RD, Fogel ML, Berry JA (1993) Photosynthetic fractionation of the stable isotopes of oxygen and carbon. Plant Physiology 101, 37–47. Hayes JM (2001) Fractionation of carbon and hydrogen isotopes in biosynthetic processes. In: Stable Isotope Geochemistry (eds Valley JW, Cole DR). Mineral. Soc. Amer., Washington, DC, pp. 225–277. Hinga KR, Arthur MA, Pilson MEQ, Whitaker D (1994) Carbon isotope fractionation by marine phytoplankton in culture: the effects of CO2 concentration, pH, temperature, and species. Global Biogeochemical Cycles 8, 91–102. Hopkinson BM, Dupont CL, Allen AE, Morel FM (2011) Efficiency of the CO2-concentrating mechanism in diatoms. Proceedings of the National Academy of Sciences 108, 3830–3837. Horken K, Tabita FR (1999) Closely related form I ribulose bisphosphate carboxylase/oxygenase molecules that possess different CO2/O2 substrate specificities. Archives of Biochemistry and Biophysics 361, 183–194. Johnston AM (1996) The effect of environmental variables on 13C discrimination by two marine phytoplankton. Marine EcologyProgress Series 132, 257–263. Kukert H, Riebesell U (1998) Phytoplankton carbon isotope fractionation during a diatom spring bloom in a Norwegian fjord. Marine Ecology-Progress Series 173, 127–137. Li Z-H, Jin H-W, Pang G-L, Zhang H-L (2010) Characters of phytoplankton community structure in coastal waters of middlenorthern Zhejiang province in Spring and Summer, 2009. Marine Fisheries 32, 401–410.

Loubere P, Bennett S (2008) Southern Ocean biogeochemical impact on the tropical ocean: stable isotope records from the Pacific for the past 25,000 years. Global and Planetary Change 63, 333–340. Maberly SC, Raven JA, Johnson AM (1992) Discrimination between 12C and 13C by marine plants. Oecologia 91, 481–492. Millero FJ, Sohn ML (1992) Chemical Oceanography. CRC Press, Ann Arbor, MI. Mook WG, Bommerson JC, Staverman WH (1974) Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide. Earth and Planetary Science Letters 22, 169–176. Park H, Song B, Morel FM (2007) Diversity of the cadmiumcontaining carbonic anhydrase in marine diatoms and natural waters. Environmental Microbiology 9, 403–413. Popp BN, Laws EA, Bidigare RR, Dore JE, Hanson KL, Wakeham SG (1998) Effect of phytoplankton cell geometry on carbon isotopic fractionation. Geochimica et Cosmochimica Acta 62, 69–77. Post DM (2002) Using stable isotopes to estimate trophic position: models, methods, and assumptions. Ecology 83, 703–718. Raven JA, Cockell CS, De La Rocha CL (2008) The evolution of inorganic carbon concentrating mechanisms in photosynthesis. Philosophical Transactions of the Royal Society B: Biological 11 Sciences 363, 2641–2650. Raven JA, Johnston AM, Kubler JE, Korb RE, Mcinroy SG, Handley LL, Scrimgeour CM, Walker DI, Beardall J, Vanderklift M, Fredriksen S, Dunton KH (2002) Mechanistic interpretation of carbon isotope discrimination by marine macroalgae and seagrasses. Functional Plant Biology 29, 355–378. Rayner PJ, Law RM, Allison CE, Francey RJ, Trudinger CM, Pickett-Heaps C (2008) Interannual variability of the global carbon cycle (1992-2005) inferred by inversion of atmospheric CO2 and delta(CO2)-C-13 measurements. Global Biogeochemical Cycles 22, 12. Read BA, Tabita FR (1994) High substrate specificity factor ribulose bisphosphate carboxylase/oxygenase from eukaryotic marine algae and properties of recombinant cyanobacterial Rubsco containing “algal” residue modifications. Archives of Biochemistry and Biophysics 312, 210–218. Riebesell U, Burkhardt S, Dauelsberg A, Kroon B (2000) Carbon isotope fractionation by a marine diatom: dependence on the growth-rate-limiting resource. Marine Ecology-Progress Series 193, 295–303. Roberts K, Granum E, Leegood RC, Raven JA (2007) Carbon acquisition by diatoms. Photosynthesis Research 93, 79–88. Robinson JJ, Scott KM, Swanson ST, O’leary MH, Horken K, Tabita FR, Cavanaugh CM (2003) Kinetic isotope effect and characterization of form II RubisCO from the chemoautotrophic endosymbionts of the hydrothermal vent tubeworm Riftia pachyptila. Limnology and Oceanography 48, 48–54. Roeske CA, O’leary MH (1984) Carbon isotope effects on the enzyme-catalyzed carboxylation of ribulose bisphosphate. Biochemistry 23, 6275–6284. Roeske CA, O’leary MH (1985) Carbon isotope effect on carboxylation of ribulose bisphosphate catalyzed by ribulosebisphosphate carboxylase from Rhodospirillum rubrum. Biochemistry 24, 1603–1607. Rost B, Riebesell U, Burkhardt S, S€ ultemeyer D (2003) Carbon acquisition of bloom-forming marine phytoplankton. Limnology and Oceanography 48, 55–67. Scott KM (2003) A d13C-based carbon flux model for the hydrothermal vent chemoautotrophic symbiosis Riftia

© 2014 John Wiley & Sons Ltd

Isotope discrimination by diatom RubisCO pachyptila predicts sizeable CO2 gradients at the host-symbiont interface. Environmental Microbiology 5, 424–432. Scott KM, Lu X, Cavanaugh CM, Liu J (2004a) Optimal methods for estimating kinetic isotope effects from different forms of the Rayleigh distillation equation. Geochimica et Cosmochimica Acta 68, 433–442. Scott KM, Schwedock J, Schrag DP, Cavanaugh CM (2004b) Influence of form IA RubisCO and environmental dissolved inorganic carbon on the d13C of the clam-bacterial chemoautotrophic symbiosis Solemya velum. Environmental Microbiology 6, 1210–1219. Scott KM, Henn-Sax M, Longo D, Cavanaugh CM (2007) Kinetic isotope effect and biochemical characterization of form IA RubisCO from the marine cyanobacterium Prochlorococcus marinus MIT9313. Limnology and Oceanography 52, 2199– 2204. Sinetova MA, Kupriyanova EV, Markelova AG, Allakhverdiev SI, Pronina NA (2012) Identification and functional role of carbonic anhydrase Cah3 in thylakoid membranes of pyrenoid of Chlamydomonas reinhardtii. Biochimica et Biophysica Acta 1817, 1248–1255. Sorhannus U, Fox MG (2012) Phylogenetic analyses of a combined data set suggest that the Attheya lineage is the closest

© 2014 John Wiley & Sons Ltd

11

living relative of the pennate diatoms (Bacillariophyceae). Protist 163, 252–262. Tabita FR, Hanson TE, Li H, Satagopan S, Singh J, Chan S (2007) Function, structure, and evolution of the RubisCO-like proteins and their RubisCO homologs. Microbiology and Molecular Biology Reviews 71, 576–599. Talavera G, Castresana J (2007) Improvement of phylogenies after removing divergent and ambiguously aligned blocks from protein sequence alignments. Systematic Biology 56, 564–577. Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S (2011) MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Molecular Biology and Evolution 28, 2731– 2739. Tcherkez GGB, Farquhar GD, Andrews TJ (2006) Despite slow catalysis and confused substrate specificity, all ribulose bisphosphate carboxylases may be nearly perfectly optimized. Proceedings of the National Academy of Sciences 103, 7246– 7251. Trimborn S, Wolf-Gladrow D, Richter K, Rost B (2009) The effect of pCO2 on carbon acquisition and intracellular assimilation in four marine diatoms. Journal of Experimental Marine Biology and Ecology 376, 26–36.