Ecological Aspects of Carbon and Nitrogen Isotope Ratios of Cyanobacteria

Plankton Benthos Res 7(3): 135–145, 2012 Plankton & Benthos Research © The Plankton Society of Japan

Ecological aspects of carbon and nitrogen isotope ratios of cyanobacteria

1, 2 2 3 3 EITARO WADA *, KAORI OHKI , SHINYA YOSHIKAWA , PATRICK L. PARKER , CHASE VAN BAALEN , 4 1 1 GENKI I. MATSUMOTO , MAKI NOGUCHI AITA & TOSHIRO SAINO

1 Environmental Biogeochemical Cycle Research Program, Research Institute for Global Change, Japan Agency for Marine- Earth Science and Technology (JAMSTEC), 3173–25 Showa-machi, Kanazawa-ku, Yokohama, Kanagawa 236–0001, Japan 2 Department of Marine Bioscience, Fukui Prefectural University, 1–1 Gakuenncho, Obama, Fukui 917–0003, Japan 3 Marine Science Institute, The University of Texas, 750 Channel View Drive Port Aransas, TX 78373, U.S.A. 4 Department of Environmental Studies, School of Social Information Studies, Otsuma Women’s University, 2–7–1 Karakida, Tama, Tokyo 206–8540, Japan Received 5 January 2012; Accepted 28 May 2012

Abstract: Variation in the δ13C of a cultured marine cyanobacterium (Agmenellum quadruplicatum, strain PR-6) is 13 described with an emphasis on the relationships between growth rate, pCO2, and δ C. An average nitrogen isotope fractionation of 1.0017 was obtained during N2 ﬁxation by cultured marine cyanobacteria. This result suggests that 15 nitrogen supply via N2 ﬁxation may be characterized by a low δ N, down to ca. －2‰ in marine environments, where the average δ15N is higher (at ca. 6‰) for major inorganic nitrogenous compounds such as nitrate. The natural abundances of nitrogen and carbon isotope ratios for cyanobacteria collected from the McMurdo Dry Valleys, Antarctica, were characterized by extremely low δ15N and widely ranging δ13C. Summarizing these and other data obtained, an isotopic map of cyanobacteria and other plankton in aquatic ecosystems was constructed and the ecological implications are discussed. Our results suggest that the δ13C of marine phytoplankton is positively correlated with sea surface temperature and can be a useful parameter for estimating in situ growth rates in the open ocean.

Key words: Antarctica, C and N isotope ratios, cyanobacteria, marine plankton, N2 ﬁxation

1991), depending on temperature, pH (which affects the Introduction HCO3–CO2 isotope exchange equilibrium), and carbon Stable isotope studies can be useful for understanding availability. Based on latitudinal, hemispheric, and polar nutrient cycling, as well as the physiological nature of mi- discrepancies in plankton δ13C, Rau et al. (1982) and Goer- croorganisms in aquatic ecosystems. In photosynthetic or- icke & Fry (1994) emphasized that geographic differences ganisms, carbon isotope fractionation is generally recog- in kinetic isotope effects associated with plankton biosyn- nized to occur during photosynthetic carbon ﬁxation. In thesis and metabolism were important factors causing lati- his comprehensive review, O’Leary (1981) demonstrated tudinal variation in plankton carbon isotope abundance. that the carbon isotope effect in the C3 pathway is large In a continuous culture of Chlamydomonas reinhardtii, during the carboxylation step. However, the diffusion of Takahashi et al. (1991) found a linear relationship between

CO2 into leaves is partially rate-limiting and this serves to the growth rate constant and the carbon isotope fraction- reduce in vivo fractionation, suggesting a possible correla- ation factor under light- and nitrogen-limited conditions. tion between the carbon isotope effect and the growth The same trend was also reported for nitrate-uptake pro- physiology of marine phytoplankton. The CO2 content of cesses in the marine diatom Phaeodactylum tricornutum supplied gas has a major effect on the δ13C of microalgae (Wada & Hattori 1978). Because information on the (Mizutani & Wada 1982). δ13C in marine plankton is growth physiology of phytoplankton and epibenthic algae thought to range from －9 to －30‰ (Wada & Hattori can likely be obtained by measuring their C and N stable isotope ratios, Wada & Hattori (1991) created a δ15N–δ13C * Corresponding author: Eitaro Wada; E-mail, [email protected] map of marine phytoplankton. 136 E. Wada et al.

In the open ocean, the nitrogen isotope ratios of plank- Materials and Methods ton are closely correlated with the chemical forms of inorganic nitrogen used for primary production. The level of Cultures δ15N in Trichodesmium sp. is low compared to those of plankton commonly found in other areas, supporting the Agmenellum quadruplicatum, strain PR-6 (Synechococus idea that nitrogen is supplied to this community via molec- sp. ATCC27264/PCC7002), was kept in the laboratory of ular nitrogen fixation (Minagawa & Wada 1986, Saino & one of the authors (Van Baalen) in an axenic form. The or- Hattori 1987, Carpenter et al. 1997). The aerobic nitrogen ganism was grown in modified ASP-2 medium (Van fixation ability of Trichodesmium has been confirmed Baalen, 1962) in test tubes (100 ml) or Erlenmeyer flasks 13 using isolated strains (Ohki et al. 1986, 1988, Chen et al. (1 L) with continuous bubbling of 1% CO2 (δ C PDB= 1996, Ohki 2008, Finzi-Hart et al. 2009). However, nitro- －36.64‰) and illuminated with fluorescent lamps 10 cm gen fixation sometimes occurs at low rates relative to the from the growth apparatus (50–60 μmol photons m－2 sec－1). total nitrogen budget (Ohki et al. 1991, Capone et al. 1997). To control the growth rate, light intensity was altered using Unicellular diazotrophic cyanobacteria contribute signifi- black mesh cloth to cover the test tubes. Growth of the liq- cantly to nitrogen cycling in the ocean (Zehr et al. 2001), uid cultures was followed turbidimetrically and the spe- although δ15N data for these unicellular species are not yet cific growth rate constant (μ) was estimated using the available. method of Kratz and Myers (1955), where μ=0.301 corre- Cyanobacteria are distributed across a wide range of sponds to a generation time of 1.00 d. physicochemical conditions, such as pH, redox potential, Cells for isotopic studies were harvested at the late loga- and temperature, and play important roles as primary pro- rithmic phase or early stationary phase for test tube cul- ducers in a variety of natural ecosystems. In the McMurdo tures. In large-scale cultures, cells were harvested at inter- Dry Valleys, Victoria Land, Antarctica, there are a variety vals and collected on Whatman type-C glass fiber filters of saline lakes and ponds that host undescribed types of preheated to 400°C for 5 h. Filters with residues were cyanobacteria with the lowest known δ15N values in the treated with 0.1 N HCl and stored in a refrigerator until biosphere. In the saline lakes and ponds the cyanobacteria mass spectrometric analysis. use nitrate supplied via precipitation. Because of the very Diazotrophs low humidity in the McMurdo Dry Valleys, the precipi- tated nitrate is present as a powdered crystal of NaNO3, Trichodesmium spp. NIBB1067 was isolated and cul- which is the nitrogen source used for growth by the cyano- tured in a synthetic medium as described in Ohki & Fujita bacteria in the saline lakes and ponds in this area. The (1982) and Ohki et al. (1986). Five unicellular and one fila- 15 δ N of sodium nitrate (NaNO3), which can be as low as mentous marine cyanobacteria isolated from the sea －20‰, and nitrogen isotope fractionation during nitrate around Singapore (Ohki et al. 2008) and Crocosphaera uptake are possible causes of these low δ15N values (Wada watsonii WH8501 (Waterbury and Pippka 1989) provided et al. 1981, Matsumoto et al. 1987, Matsumoto et al. 1993). by Drs. Waterbury and Dyhman of Woods Hole Oceano- Hage et al. (2007) isotopically characterized coastal pond- graphic Institution, USA, were maintained in a liquid mod- derived organic matter in the McMurdo Dry Valleys. Re- ified Aquil medium without combined nitrogen (Ohki et al. cently, Hopkins et al. (2009) reported isotopic evidence for 1986). Two freshwater heterocystous cyanobacteria in sec- the provenance and turnover of organic carbon by soil mi- tions IV and V (Nostoc sp. PCC7120 and Fischerella croorganisms in the McMurdo Dry Valleys. sp1ATCC29114, respectively) were cultured in BG0 (Rip- In this study, we investigated the effects of cyanobacte- pka & Herdman 1992), and used as references. Cultures rial growth rate on carbon isotope fractionation in a ma- were kept under fluorescent lights (daylight type) at rine cyanobacterium (Agmenellum quadruplicatum; PR-6). 10 μmol photons m－2 sec－1 (12 : 12 L : D) at 26°C (Ohki et Marine diazotrophs were also investigated to elucidate the al. 2008). Several milligrams of cells were collected, cen- role of nitrogen isotope fractionation in molecular nitrogen trifuged at 3,000×g for 4 min at 4°C, and lyophilized for fixation by various kinds of marine cyanobacteria. Cyano- stable isotope measurements. bacteria collected from saline lakes and ponds in the Mc- Cyanobacteria from the McMurdo Dry Valleys, Ant- Murdo Dry Valleys, Antarctica, were also studied in detail, arctica with an emphasis on cyanobacterial growth physiology. Finally, we summarized the characteristics of cyanobacte- Observations of the McMurdo Dry Valleys area during ria on a δ15N vs. δ13C map and discuss some ecological im- the 1980–1981 field season were reported in detail by plications of algal isotope ratios based on our data in com- Wada et al. (1984). Lake Vanda, several ponds in the Laby- bination with previously reported data (Minagawa & Wada rinth, Lake Bonney, and other lakes were visited during 1986, Wada & Hattori 1991, Yoshioka et al. 1994, Aita et the season. Epibenthic cyanobacteria were collected from al. 2011). unnamed saline ponds in the Labyrinth, near the terminus of the Wright Upper Glacier, and from Lake Vanda and its surroundings. Samples were also collected from the Taylor C and N isotope ratios of cyanobacteria 137

Fig. 1. Locations of sampling sites in the McMurdo Dry Valleys of southern Victoria Land, Antarctica (revised from Matsu- moto et al. 1990). Dotted areas are ice-free. South side of Lake Vanda (77°31′0″S, 161°40′0″E): V1, V2, V3, V4, V5, V6, V7, V8, V9, V10, and V11. Surroundings of Lake Canopus (77°33′00.0″S, 161°31′00.0″E): C2 (dried cyanobacteria), C3 (wet cyanobacteria), C4 (wet cyanobacteria), C8 (wet cyanobacteria), and C9 (dried cyanobacteria). Don Juan Pond (77°33′55″S, 161°11′26″E): D1 (dried cyanobacteria), D2 (dried cyanobacteria), and D3 (dried cyanobacteria). Unnamed ponds in the Labyrinth (77°33′S, 160°50′E): L11, L12, L13, L14, and L20. Lake Bonney (77°43′S, 162°22′E): B1 (cyanobacteria in the meltwater stream at the east side of the lake), B2, and B3 (cyanobacteria in the meltwater stream from Hughes Glacier).

Valley (Fig. 1). Isotope analyses Samples were frozen and brought back to Japan. Prelim- inary results from these samples were reported by Wada et The samples collected from Antarctica were analyzed al. (1981). Sterol distribution in the saline ponds can be using an RMU-6R mass spectrometer (Hitachi) fitted with found in Matsumoto et al. (1982). The geochemical fea- a double collector for ratiometry. The cultured diazotrophs tures of the McMurdo Dry Valley lakes and ponds were re- were measured by Flash 2000, CoFloIV, Delta V (Thermo viewed by Matsumoto (1993). Fisher Scientific) in SI Science Co., Ltd., Japan (Aita et al. To identify cyanobacteria, small aliquots of freeze-dried 2011). The isotopic data are presented in terms of δ13C as samples (5–50 mg) were taken, and DNA was extracted the per mil deviation relative to the PDB isotopic standard. using a FastDNA Spin kit (MP Biomedicals, Solen, OH, For δ15N, atmospheric nitrogen was used as a standard. USA). Most of the 16S rRNA gene fragments were ampli- The precisions of δ13C and δ15N measurements were within fied using cyanobacteria-specific primers as follows: 8F 0.2‰. (forward, Wilmotte et al. 1993) and PISTE-Cyano-R (re- 13 15 3 verse, CTCTGTGCCAAGGTATC, Harverkamp, unpub- ddC or N (‰ ) = (RRsample / standard－ 1) ×10 (1) lished data). The conditions for polymerase chain reaction (PCR) were the same as those used in Ohki et al. (2008). where R is 13C/12C or 15N/14N, respectively. 13 The PCR products were ligated into a pGEM-T vector The δ C values of PR-6 are reported relative to the CO2 (Promega, WI, USA), and transformed into Escherichia used as the carbon source. coli cells to construct clone libraries. Clones containing an Carbon isotope discrimination (Δδ13C) is generally ex- insert were selected and sequenced after the plasmid was pressed as a difference in the δ13C value between source purified. Homology-based searches of the DNA Data Bank and product (Takahashi et al. 1991): of Japan (DDBJ) were performed using the BLAST pro- 13 13 gram on the DDBJ website (http://www.ddbj.nig.ac.jp/ δ C (source)－δ C (product) ∆d 13C= (2) index-j.html). 1+δ13 C (source) /1000

Spontaneous discrimination factor at biomass Nt1 and Nt2 138 E. Wada et al.

correlation with growth rate; that is, the δ13C of the cells decreased with decreasing growth rate. The other group, shown within the dotted circle in Fig. 2, which had rather low discrimination factors, were collected in the late logarithmic phase or early stationary phase. Under those conditions, cell numbers in the incubation tube were maximal, meaning that high photosynthetic activity was rapidly con- suming dissolved carbon dioxide. Diazotrophs The nitrogen isotope ratios (δ15N) of cultured cyanobac-

teria during N2 ﬁxation ranged from －1.5 to －3.0‰ relative to atmospheric N2, with an average value of －1.74‰, corresponding to a fractionation factor of 1.0017 (Table 1). These values were constant regardless of algal species and

were comparable to those obtained for other N2-fixing microorganisms, including Azotobacter. Fractionation during biological molecular nitrogen fixation was almost identical Fig. 2. Summary of carbon isotope discrimination factor among these organisms. (Δδ13C) during the growth of Agmenellum quadruplicatum. Cul- The partial rDNA sequence of cyanobacteria in the Mc- tures were constantly bubbled with 1.0% CO2 in air at 36.5°C. FL-series were conducted using 1-L Erlenmeyer flasks. Instanta- Murdo Dry Valleys, Antarctica 13 neous discrimination factor (Δδ C) between time t1 and time t2 Diazotrophic cyanobacteria closely related to the genera was calculated using a mass balance equation. TT-series were Leptolyngbya, Oacillatoria, and Phormidium (N2-fixing cultured in a 100-mL test tube and harvested in the early station- under anaerobic conditions, cf. Rippka & Herdman 1992) ary phase when algal cell density and photosynthetic activity were detected by partial sequencing of PCR-amplified 16S were high. The observed low Δδ13C in the FL-series might have rDNA using DNA extracted from the samples collected resulted from CO limitation with rather high growth and very 2 from Antarctic saline ponds and lakes in the McMurdo high algal cell density during the late logarithmic phase. Such Dry Valleys. conditions were sometimes observed when bubbling of CO2 en- riched air became insufficient during laboratory culturing. Natural C and N isotope abundances of cyanobacteria in McMurdo Dry Valleys in the flask cultures was estimated by the following isotope The natural C and N abundances of cyanobacteria col- mass balance equation: lected from saline ponds and lakes in McMurdo Dry Val- leys were precisely measured. δ15N values divided samples δ13C × f +δ13 C (1－ f )= δ 13 C (3) into two groups with average values of －10 and －50‰, Nt12NΔt Nt respectively (Fig. 3 and Appendix). According to Wada et 15 where f=Nt1/Nt2 and Δt=t2－t1. al. (1981), the δ N of soil nitrate in McMurdo Dry Valleys 13 Equation (3) can be re-arranged to obtain δ CNΔt, which is quite low, from －24 to －11‰. This can be considered a is the instantaneous discrimination factor during the inter- result of the characteristic origins of nitrate associated val between t1 and t2. This was done for the FL-1 and FL-2 with atmospheric precipitation, including the production of 15 incubation series in Fig. 2. NOx by auroral activity. The former group (δ N of －10‰) reflects source nitrate δ15N, whereas the latter (low δ15N) group probably arose under conditions of high nitrate con- Results centration and low light intensity, with a significant occurrence of nitrogen isotope fractionation (up to 1.04) during Agmenellum quadruplicatum (PR-6) nitrate assimilation. In fact, the occurrence of a large nitro- Cyanobacterial cells with different growth rates under gen isotope fractionation was observed for several saline different light intensities exhibited marked variation in ponds in the Labyrinth (Table 2). On one hand, for low 13 13 － δ C values (Fig. 2). The δ C of cells ranged from －19.1 to NO3 concentrations down to 1 μM, no nitrogen isotope －10.7‰, relative to the CO2 used as the substrate. These discrimination is expected during nitrate assimilation. values correspond to discrimination factors of 19.8 and Consequently the δ15N of cyanobacteria in such ponds be- 11.1, respectively. The maximum discrimination factor was come close to the δ15N of nitrate (－20‰) used for growth. observed in cells grown under low light intensities with However, the δ13C was highly variable, ranging from －25 13 low growth rates. The variation in δ C of the cells was di- to －3‰. High values resulted from the limitation of CO2 vided into two groups: For one group there was a negative supply during growth. Algal mats are well known to show C and N isotope ratios of cyanobacteria 139

Table 1. Nitrogen isotope ratios of N2-ﬁxing cyanobacteria during N2 ﬁxation.

Organisms δ15N (‰) References

Azotobacter Azotobacter agile －3.5 Hoering & Ford (1960) Azotobacter chroococcum －1.9 ibid. Azotobacter vinelandii －4.1 ibid. Blue green Anabaenacylindrica －3.0 Wada & Hattori (1974) algae －0.6 Minagawa & Wada (1986) Trichodesmium spp. －2.2 This work －1.3 Carpenter et al. (1997) －3.7 Carpenter et al. (1997) Gloeothece. sp. 68DGA (unicellular) －1.5 strain: Ohki et al. (2008) Gloeothece. sp. 11DG (unicellular) －1.6 strain: Ohki et al. (2008) Gloeothece. sp. 30D1 (unicellular) －1.5 strain: Ohki et al. (2008) Gloeothece. sp. 20B5 (unicellular) －1.4 strain: Ohki et al. (2008) Gloeothece. sp. 38U6 (unicellular) －1.8 strain: Ohki et al. (2008) Crocosphaera watsonii WH8501 (unicellular) －2.2 strain: Waterbury et al. (1989) Lyngbya sp. 10B (Filamentous) －1.7 Nostoc sp. PCC7120 －1.8 Fischerella sp1ATCC29114 －2.0 Trichodesmium spp. －1.3 and －3.7 Reported by Carpenter et al. (1997)

mean －1.7 Only for cyanobacteria except data by Carpenter et al. (1997)

Fig. 3. δ15N–δ13C map for cyanobacteria in the McMurdo Dry Valleys, Antarctica. For details, see Fig. 1 and Appendix.

F1 F2 13  → CO22← CO → Organic-C such high δ C values. F3

The ﬁrst step is CO transport into the cell through bidi- Discussion 2 rectional passive diffusion (F1 and F3), and the next is assimilation (F2) into organic carbon by ribulose bisphos- Theoretical consideration of steady-state C/N kinetic phate carboxylase (RuBPCase). isotope effects Under the steady state condition during photosynthetic

Isotope fractionation during photosynthetic CO2 ﬁxation carbon ﬁxation, we obtain and nitrate assimilation can be described in the following way (Farquhar et al. 1989, Takahashi et al. 1991). ∆δ 13C=b +× ( a－ b ) F2 / F1 (4) 140 E. Wada et al.

Table 2. Nitrogen isotope fractionation during cyanobacterial growth in saline ponds of the Labyrinth in the McMurdo Dry Valleys, Antarctica. Nitrogen isotope fractionation factors up to 1.04 were observed during nitrate-uptake processes under high nitrate concentrations.

NH+－N NO－+NH+ δ15N (NH+) δ15N (NO－) δ15N (cyanobacteria) δ13C (cyanobacteria) Pond 4 3 4 4 3 (mgN L－1) (mgN L－1) (‰) (‰) (‰) (‰)

L-11 4.5 455 －22.4 －11.6 －50.0 －4.55 L-12 10.3 1,370 －21.2 －10.0 －51.4 －6.55 L-13 6.6 1,290 n.d. －1.6 －52.7 －7.76 L-14 8.3 1,370 －8.0 －5.6 －63.4 －13.30 S-1 Valley 14.2 14.2 +2.0 －13.2 －19.8 －1.50 where a and b are the discrimination factors (DF) associ- PR-6 in the present study also suggests that carbon isotope ated with the processes of CO2 diffusion (F1) and carbo- fractionation is regulated by the kinetics of photosynthetic xylation reaction (F2), respectively (Takahashi et al. 1991). carbon ﬁxation pathways. The differential or instantaneous For this simple model, we can obtain DF of 4.4 and 30 for carbon isotope discrimination factor decreased with an in- the diffusion process and carboxylation reaction, respec- crease in its growth rate, similar to Chlamydomonas rein- tively. Then we obtain hardtii Dangeard (IAM C-238) described above. We can conclude that the δ13C content of PR-6 decreases with in- Δ+×δ 13C = 30 (4.4－ 30) F2/ F1 (5) creasing growth rate, as indicated by the shaded broad line in Fig. 2. In contrast, the low Δδ13C for the data within the

Here we predict that the quotient F2/F1 in the above equa- circle in Fig. 2 strongly suggests that the effect of CO2 lim- tion is proportional to μ, because F1 seems to be indepen- itation on algal photosynthetic activity lowered the appar- dent of μ and kept constant under constant ambient pCO2. ent fractionation factor at the rather high growth rates of Thus, a constant F1 flux implies a negative relationship dense populations within the small incubation tubes. In between Δδ13C and either F2 or μ. In fact, during algal summary, we conclude that algal δ13C is governed by two photosynthesis, ΔD has a negative relationship with the major factors: growth rate and the supply of CO2 as esti- growth rate constant μ (Takahashi et al. 1991). The follow- mated by the above eq. (5). ing equation was obtained by Takahashi et al. (1991) for Diazotrophs Chlamydomonas reinhardtii Dangeard (IAM C-238): Molecular nitrogen has a very strong triple bond be- ∆δµ13C=－－ 35.4 += 25.3 (r 0.92) (6) tween nitrogen atoms. Theoretically, the cleavage of nitrogen atoms under a reversible reaction results in a very The nitrogen isotopic fractionation factor is inversely cor- large nitrogen isotope fractionation. However, the ob- related with the growth rate constant in the assimilation of served fractionation was small for biological molecular ni- nitrate by the marine diatom Phaeodactylum tricornutum. trogen fixation, ca. 1.002. Such a small nitrogen isotope The nitrogen isotope fractionation factor is 1.001 at high fractionation factor strongly suggests that triple-bond growth rates of P. tricornutum, and goes up to 1.023 under cleavage takes place suddenly under irreversible conditions light-limited conditions (Wada & Hattori 1978). In view of between substrate N2 and an activated substrate. Judging these facts, variation in cellular δ13C was examined for the from the present results and from data in the literature, we cyanobacterium PR-6. conclude that isotope fractionation during biological molecular nitrogen fixation is 1.002 in general. This means Agmenellum quadruplicatum (PR-6) that the δ15N supplied via molecular nitrogen fixation is ca. The enzymatic carboxylation reaction is the primary －2‰, a fairly low value in marine environments and factor contributing to carbon isotope fractionation during eutrophic lakes, where the N2-fixing process is easily dis- photosynthesis (Christeller et al. 1976, Estep et al. 1978). tinguished by the nitrogen isotope ratio and where N2-fix- The temperature-dependent isotope exchange equilibrium ing cyanobacteria exhibit low δ15N and high δ13C. These between CO2 and bicarbonate, pCO2, CO2 diffusion, and results support the findings of other studies that have dem- diffusion/carboxylation partitioning (F2/F1 in the eq. (5)) onstrated the significance of molecular nitrogen fixation in relating to the growth physiology of organisms are also the western North Pacific Ocean (Saino & Hattori 1987, principal components governing the magnitude of carbon Wada & Hattori 1991). isotope fractionation (O’Leary 1981). Of these factors, the Epibenthic cyanobacteria from McMurdo Dry Valleys growth physiological condition is most likely to cause fluc- tuation or deviation in the 13C content of phytoplankton in The carbon and nitrogen isotope ratios of epibenthic cy- the surface waters of the ocean. The variation in δ13C in anobacteria can be explained by the kinetic isotope effects C and N isotope ratios of cyanobacteria 141 described above during cyanobacterial growth in Mc- Murdo Dry Valleys. The nitrogen and carbon isotope ratios of microalgae closely depend on the isotope ratios of the substrate used for their growth as well as on their uptake kinetics. Under substrate-limiting conditions, isotope fractionation factors become close to 1.000 because the diffusion of the substrate into algal cells is the rate-limiting step during the assimilation of inorganic nitrogen, whereas large fractionation factors can occur when enzymatic processes limit the overall reaction under low-light conditions with slow growth rates. As reported by Wada et al. (1984), the δ15N of sodium nitrate in the McMurdo Dry Valleys is as low as －20‰. Species of cyanobacteria found in the McMurdo Dry Valleys by Matsumoto et al. (1993, 1996) included Lyngbya murrayi, Oscillatoria priestley, and Phormidium priestley. According to Matsumoto, who par- ticipated in field surveys in this area from 1976 through 1985, our samples are consistent with these organisms (GI Matsumoto, pers. comm.). The results of our 16S rDNA sequence analysis support the observations of Matsumoto et al. (1993, 1996). In the present work, we detected two algal groups based on nitrogen isotope ratios, i.e., －20‰ and －80 to －50‰ Fig. 4. Relationship between the δ13C of total organic carbon (Fig. 3). The former group might grow under substrate- in pelagic plankton and surface water temperature (SST). Data limited conditions caused by low nitrate concentrations sources: Wada and Hattori (1991), Aita et al. (2011). Unpublished and/or the aggregation of algal mats at the bottom of saline 15 data collected from the Pacific Coast of Japan are included in the ponds. However, very low δ N values could also occur figure by their sampling location name. when algal growth is limited by light intensity under high nitrate concentrations, up to 1,000 mg nitrate-N L－1. In such cases, we expect the occurrence of a high nitrogen isotope fractionation factor (up to ca. 1.04) during nitrate concentration on nitrate uptake as measured in shipboard assimilation processes. In McMurdo Dry Valleys, there are experiments, Smith (2010) found evidence of greater tem- several saline ponds with the conditions listed in Table 2. perature sensitivity than that reported by Eppley (1972) for laboratory measurements of growth rate. The lowest δ13C Relationship between the δ13C of marine plankton and values (ca. －30‰) were those of plankton from high-lati- sea surface temperature (SST) tude areas (Rau et al. 1982, Goericke & Fry 1994). We The reddish brown Trichodesmium erythraeum could thus expect to see a positive relationship between the (Marumo & Asaoka 1974) has the highest δ13C among the growth rate constant and either δ13C in phytoplankton pelagic plankton samples documented in the literature (POC) or SST. In fact, we obtained a linear relationship for (Fig. 4). Nitrogen supply via molecular nitrogen fixation is phytoplankton: δ13C=0.25 SST－26.2 (r2=0.59), as indi- responsible for the low δ15N values of Trichodesmium cated in Fig. 4. This linear relationship observed in the (Wada 1980, Minagawa & Wada 1986). Therefore, Tricho- western North Pacific off Japan is quite similar to that re- desmium holds a unique position among marine phyto- ported by Goericke and Fry (1994), whose linear regres- 13 13 plankton with regard to its carbon and nitrogen isotopic sion of δ CPOC on SST was δ CPOC=0.17 [±0.04]×SST－ composition (Fig. 4), as described by Carpenter et al. 24.6‰ (r=0.72, >5°C in the northern hemisphere). (1997). Therefore, phytoplankton δ13C is a possible parameter The relationship between δ13C and growth rate constant for deducing in situ growth rate constants in the marine was examined, based on all marine phytoplankton δ13C environment. Thus, the variability in δ13C content among data collected from the Pacific Coast off Japan (Fig. 4). various types of particulate organic matter in certain ocean 13 Generally, phytoplankton δ C depends on pCO2, growth waters suggests that stable carbon isotope studies could rate, SST, nutrient concentration, and the cell geometry of provide unique information about the growth physiology each species (Popp et al. 1998). According to Eppley as well as the carbon dynamics of the plankton commu- (1972), seawater temperature and nutrient concentrations nity. are two major factors affecting phytoplankton growth. By accounting for the combined effects of temperature and 142 E. Wada et al.

Fig. 5. δ15N–δ13C map of marine phytoplankton and cyanobacteria with an emphasis on diazotrophs. (a) δ15N–δ13C map of marine phytoplankton (×; Wada & Hattori, 1991) and Trichodesmium spp. (□; this study), cyanobacteria in Lake Suwa, Japan (●), cyanobacteria from the McMurdo Dry Valleys (◯), algae from human-impacted Brazilian lakes (▽), and Microcystis from Lake Barra Bonite under sewage loading from São Paulo, Brazil (▼; Wada et al. 1991). The values for marine phytoplankton and cyanobacteria from Lake Suwa almost overlap, as indicated by the shaded area. The cyanobacteria from the McMurdo Dry Valleys had extremely low δ15N and a wide range of δ13C values. (b) δ15N–δ13C map of cyanobacteria in Lake Suwa, Japan. Nu- merical values in circles correspond to the months when samples were collected. Data were obtained from Yoshioka et al. (1994). (c) δ15N–δ13C map of marine phytoplankton from the western North Paciﬁc, following Wada and Hattori 1991. Data from the Oyashio current area are from Aita et al. (2011).

Blooms of Microcystis are often observed from spring to An isotopic map of cyanobacteria and plankton summer in this highly eutrophic lake. Microcystis utilizes Lake Suwa (36°03′N, 138°05′E) is a typical eutrophic nitrate from domestic sewage in spring, then use regener- shallow lake in Nagano prefecture, in central Japan, with a ated ammonium in summer. After that the N2 ﬁxing cya- surface area of 13.3 km2 and a maximum depth of 6.8 m. nobacteria Anabaena typically appears under nitrogen-de- C and N isotope ratios of cyanobacteria 143

ficient and phosphate rich conditions in August. The pat- for T. erythraeum. tern of succession in this cyanobacterial community pro- The natural abundances of cyanobacteria collected from vided a useful comparison with the regional differences Antarctic saline ponds in the McMurdo Dry Valleys were between marine phytoplankton when plotted on the δ15N– measured precisely. Their δ15N values were divided into δ13C map (Fig. 5a). two groups, with average values of －10 and －50‰, re- Data from a 1985 bloom (Yoshioka et al. 1994) are spectively. δ13C was highly variable, from －25 to －3‰. shown in Fig. 5b, together with plankton data from the pe- The algal growth physiology of these cyanobacteria is dis- lagic ocean in Fig. 5c (Wada & Hattori 1991). Furthermore, cussed on the basis of their carbon and nitrogen isotope ra- data from human-impacted lakes in tropical areas in Brazil tios. were also included in Fig. 5a to demonstrate how human An isotopic map of cyanobacteria and marine plankton activity such as dam construction and sand mining was presented that summarizes data appearing in the pres- changed δ15N and δ13C values of phytoplankton (Wada et ent work and in the literature. The isotope ratios of cyano- al. 1991). Coincidentally, the δ15N of inorganic nitrogenous bacteria in natural aquatic ecosystems have extremely compounds in Lake Suwa was quite similar to that of pe- wide ranges. Marine phytoplankton and the cyanobacteria lagic nitrate (ca. 6‰). The isotopic map of pelagic marine of Lake Suwa were divided into three groups (nitrate-rich, phytoplankton, Microcystis, and Anabaena cylindrica, the nitrate-poor, and N2-fixing communities) based on nutrient N2 fixing cyanobacteria that appear in August in Lake availability and growth rate. Suwa, Japan, can be divided into three groups (nitrate-rich, nitrate-poor, and N -fixing communities) depending on nu- 2 Funding trient availability, from the nitrate-rich spring season to nitrate poor-phosphate rich summer season (Fig. 5b). This work was supported by Grant-in-Aid for Scientific As shown in Fig. 5a, cyanobacteria and phytoplankton Research (C) number 20580208 to KO and a Grant-in-Aid were found in almost the same areas of the isotopic map, for Challenging Exploratory Research (22651007) to MNA strongly suggesting that the isotopic compositions of algae from the Japan Society for the Promotion of Science. are governed by a growth physiology common among marine algae and cyanobacteria. In the figure, a dramatic in- References crease can be observed in the δ15N values of phytoplankton in tropical lakes in Brazil affected by human impacts such Aita MN, Tadokoro K, Ogawa NO, Hyodo F, Ishii R, Smith SL, as sand mining, dam construction, and sewage loading Saino T, Kishi MJ, Saitoh S, Wada E (2011) Linear relationship (Wada et al. 1991). Compared to data for marine phyto- between carbon and nitrogen isotope ratios along simple food plankton (the shaded areas in Fig. 5a), the isotopic data for chains in marine environments. J Plankton Res 33: 1629–1642. cyanobacteria varied with habitat structure. The lowest Capone DG, Zehr JP, Paerl HW, Bergman B, Carpenter EJ value, a δ15N of －76.5‰, was obtained from cyanobacteria (1997) Trichodesmium, a globally significant marine cyano- collected at Lake Bonney, in the McMurdo Dry Valleys bacterium. Science 276: 1221–1229. (Fig. 3). Carpenter EJ, Harvey HR, Fry B, Capone DC (1997) Biogeo- chemical tracers of the marine cyanobacterium Trichodes- mium. Deep-Sea Res I: Oceanographic Research Papers 44: Conclusion 127–138. The relationship between growth rates and carbon iso- Chen YB, Zehr JP, Mellon M (1996) Growth and nitrogen fixation of the diazotrophic filamentous nonheterocystous cyano- tope fractionation during photosynthesis was investigated bacterium Trichodesmium sp. IMS101 in defined media: evi- using a marine cyanobacterium (A. quadruplicatum). dence for a circadian rhythm. J Phycol 32: 916–923. Under light-limiting growing conditions, rapid growth of Christeller JT, Loing WA, Troughton JH (1976) Isotope discrimi- the organism was accompanied by low isotope fraction- nation by ribulose 1,5-diphosphatecarboxylase. Plant Physiol ation factors (μ=1.6, α=1.017; μ=0.20, α=1.024); that is, 57: 580–582. δ13C was positively correlated with the growth rate. Our 13 Craig H (1957) Isotopic standards for carbon and oxygen and results also suggest that the δ C of marine phytoplankton correction factors for mass-spectrometric analysis of carbon may be useful for estimating in situ growth rates in the dioxide. Geochim Cosmochim Acta 12: 133–149. open ocean as a first approximation. We also suggest that Eppley RW (1972) Temperature and phytoplankton growth in the 13 15 the isoscape of phytoplankton δ C and δ N over the sea. Fish Bull 70: 1063–1085. oceans could help to verify marine ecosystem models by Estep MF, Tobita FR, Parker PL, Van Baalen C (1978) Carbon their distributions. isotope fractionation by ribulose-1,5-biphosphatecarboxylase Various marine N2-fixing cyanobacteria were cultured from various organisms. Plant Physiol 61: 680–687. under N2-fixing conditions to elucidate nitrogen isotope Farquhar GD, Ehlernger JR, Hubick KT (1989) Carbon isotope 15 fractionation values during N2 fixation. δ N ranged from discrimination and photosynthesis. Annu Rev Plant Physiol －2.27 to －1.48‰, with an average value of －1.74‰. The Plant Mol Biol 40: 503–537. highest δ13C among pelagic plankton (－13.6‰) was found Finzi-Hart JA, Pett-Ridge J, Eeber PK, Popa R, Fallon SJ, 144 E. Wada et al.

Gunderson T, Hutcheon ID, Nealson KH, Capone DG (2009) Ohki K, Rueter JG, Fujita Y (1986) Cultures of the pelagic cy- Fixation and fate of C and N in the cyanobacterium Trichodesmium anophytes Trichodesmium erythraeum and T. thiebautii in using nanometer-scale secondary ion mass spectrometry. Proc synthetic medium. Mar Biol 91: 9–13. Natl Acad Sci USA 106: 6345–6350. Ohki K, Rueter JG, Fujita Y (1988) Aerobic nitrogenase activity Goericke R, Fry B (1994) Variations of marine plankton δ13C measured as acetylene reduction in the marine non-heterocys-

with latitude, temperature, and dissolved CO2 in the world tous cyanobacterium Trichodesmium spp. grown under artiﬁ- ocean. Glob Biogeochem Cycl 8: 85–90. cial conditions. Mar Biol 98: 111–114. Hage MM, Uhle ME, Macko SA (2007) Biomarker and stable Ohki K, Honda D, Kumazawa S, Ho KK (2008) Morphological isotope characterization of coastal pond-derived organic mat- and phylogenetic studies on unicellular diazotrophic cyano- ter, McMurdo Dry Valleys, Antarctica. Astrobiology 7: 645– bacteria (Cyanophytes) isolated from the coastal waters 661. around Singapore. J Phycol 44: 142–151. Hoering T, Ford HT (1960) The isotope effect in the ﬁxation of Ohki K, Zehr JP, Fujita Y (1991) Trichodesmium: Establishment

nitrogen by Azotobacter. J Am Chem Soc 82: 376–378. of culture and characteristics of N2-fixation. In: Marine Pe- Hopkins DW, Sparrow AD, Gregorich EG, Elberling B, Novis P, lagic Cyanobacteria: Trichodesmium and other Diazotrophs Fraser F, Scrimgeour C, Dennis PG, Meier-Augenstein W, (eds Carpenter EJ, Capone DG, Rueter JG). Kluwer Academic Greenfield LG (2009) Isotopic evidence for the provenance Publishers, Dordrecht, pp. 307–316. and turnover of organic carbon by soil microorganisms in the O’Leary MH (1981) Carbon isotope fractionation in plants. Phy- Antarctic dry valleys. Environ Microbiol 11: 597–608. tochemistry 20: 553–567. Kratz W, Myers J (1955) Nutrition and growth of several blue- Popp B, Laws EA, Bidigare RR, Dore JE, Hanson KL, Wakeham green algae. Am J Botany 42: 282–287. SG (1998) Effect of phytoplankton cell geometry on carbon Marumo R, Asaoka O (1974) Distribution of pelagic blue-green isotopic fractionation. Geochim Cosmochim Acta 62: 69–77. algae in the North Pacific Ocean. J Oceanogr Soc Jap 30: 77– Rau GH, Sweeney ES, Kaplan IR (1982) Plankton 13C : 12C ratio 85. changes with latitude: differences between northern and Matsumoto G, Torii T, Hanya T (1982) High abundance of algal southern oceans. Deep Sea Res 29: 1035–1039. 24-ethylcholesterol in Antarctic lake sediment. Nature 299: Rippka R, Herdman M (1992) Pasteur culture collection of cya- 52–54. nobacterial Strains in axenic culture. Catalogue and taxonomic Matsumoto GI (1993) Geochemical features of the McMurdo Dry handbook, Vol. 1: catalogue of strains 1992/1993. Institut Pas- Valley lakes. In: Physical and Biogeochemical Processes in teur, Paris, France, pp. 1–103. Antarctic Lakes (Antarct. Res. Ser. 59) (eds. Green W, Fried- Saino T, Hattori A (1987) Geographical variation of the water man EI). American Geophys Union, Washington DC, pp. 95– column distribution of suspended particulate organic nitrogen 118. and its 15N abundance in the Pacific and its marginal seas. Matsumoto GI, Akiyama M, Watanuki K, Torii T. (1990) Un- Deep Sea Res 34: 807–827. usual distributions of long-chain n-alkanes and n-alkenes in Smith LS (2010) Untangling the uncertainties about combined Antarctic soil. Org Geochem 15: 403–412. effects of temperature and concentration on nutrient uptake Matsumoto GI, Ohtani S, Hirota K (1993) Biogeochemical fea- rates in the ocean. Geophys Res Lett 37, L11603, doi:10.1029/ tures of hydrocarbons in cyanobacterial mats from the Mc- 2010GL043617, 2010 Murdo Dry Valleys, Antarctica. Proc NPIR Symp Polar Biol 6: Takahashi K, Wada E, Sakamoto M (1991) Relation between car- 98–105. bon isotope discrimination and the specific growth rate of Matsumoto GI, Watanuki K, Torii T (1987) Total organic carbon green alga, Chlamydomonas reinhardtii. Jpn J Limnol 52: in ponds from the Labyrinth of Southern Victoria Land in the 105–112. Antarctic. Antarc Rec 31: 171–176. Van Baalen C (1962) Studies on marine blue-green algae. Botan- Matsumoto GI, Yamada S, Ohtani S, Broady PA, Nagashima H ica Marina 4: 129–139. (1996) Biogeochemical features of hydrocarbons in cultured Wada E (1980) Nitrogen isotope fractionation and its significance cyanobacteria and green algae from Antarctica. Proc NPIR in biogeochemical processes occurring in marine environ- Symp Polar Biol 9: 275–282. ments. In: Isotope Marine Chemistry (eds Goldberg ED, Minagawa M, Wada E (1986) Nitrogen isotope ratios of red tide Horibe Y, Saruhashi K). Uchida Rokkakuho Co., Tokyo, organisms in the East China Sea: a characterization of the bio- Japan, pp. 375–398. logical nitrogen fixation. Mar Chem 19: 245–259. Wada E, Hattori A (1978) Nitrogen isotope effects in the assimi-

Mizutani H, Wada E (1982) Effect of high atmospheric CO2 con- lation of inorganic nitrogenous compounds by marine diatoms. centration on 13C of algae. A possible cause for the average de- Geomicrobiol 1: 85–101. pletion of 13C in Precambrian reduced carbon. Origin of Life Wada E, Hattori A (1991) Nitrogen in the sea: forms, abun- 12: 377–390. dances, and rate processes. CRC Press, Boca Raton, Florida, Ohki K (2008) Intercellular localization of nitrogenase in a non- 208 pp. heterocystous cyanobacterium (Cyanophytes), Trichodesmium Wada E, Imaizumi R, Nakaya S, Torii T (1984) 15N abundance in sp. NIBB1067. J Oceanogr 64: 211–216. the Dry Valley Area, South Victoria Land, Antarctica: eco- Ohki K, Fujita Y (1982) Laboratory culture of the pelagic blue- physiological implications of microorganisms. Mem Natl Inst green alga Trichodesmium thiebautii: conditions for unialgal Polar Res Spec Iss 32: 130–139. culture. Mar Ecol Prog Ser 7: 185–190. Wada E, Lee JA, Kimura M, Koike I, Reeburgh WS, Tundisi JG, C and N isotope ratios of cyanobacteria 145

Yoshinar T, Yoshioka T, Vuuren MI (1991) Gas exchange in relationships among the cyanobacteria based on 16S rRNA se- ecosystems: framework and case studies. Jpn J Limnol 52: quences. In: Bergey’s Manual of Systematic Bacteriology. 2nd 263–281. ed. Vol. 1 (eds Boone D, Castenholz RW, Garrity GM). Wada E, Shibata R, Torii T (1981) 15N abundance in Antarctica: Springer, New York, pp. 439–514. origin of soil nitrogen and ecological implications. Nature 292: Yoshioka T, Wada E, Hayashi H (1994) A stable isotope study on 327–329. seasonal food web dynamics in a eutrophic lake. Ecology 75: Waterbury JB, Rippka R (1989) Subsection I. Order Chroococ- 835–846. cales Wettstein 1924, emend. Rippka et al., 1979. In: Bergy’s Zehr, JP, Waterbury JB, Turner PJ, Montoya JP, Omoregie E, Manual of Systematic Bacteriology, Vol. 3 (eds Staley JT, Bry- Steward GF, Hansen A, Karl DM (2001) Unicellular cyanobac-

ant MP, Pfennig N, Holt JG). Williams and Wilkins, Balti- teria ﬁx N2 in the subtropical North Paciﬁc Ocean. Nature 412: more, pp. 1728–1746. 635–637. Wilmotte AG, Van der Auwera, DeWachter R (1993) Phylogenic

Appendix. Carbon and nitrogen isotope ratios of cyanobacteria collected from the McMurdo Dry Valleys, Antarctica.

δ13C δ15N Site Sample no. Remarks (‰) (‰)

Wright Valley Lake Vanda V-A-1 －17.5 －10.5 Dried cyanobacteria under a stone Lake Vanda V-A-2 －18.5 －10.3 Under sand at 5 cm depth Lake Vanda V-A-2″ －17.7 －9.5 Under the ice at 5 cm depth Lake Vanda V-A-3 －18.1 －10.3 Degraded cyanobacteria Lake Vanda V-A-5 －7.9 －14.7 Wet cyanobacteria at the edge of the cape at 50 cm depth Lake Vanda V-A-5′ －6.1 －13.4 ibid. Lake Vanda V-A-7 －20.7 －12.3 Attached to a rock at the river mouth Lake Vanda X1 －16.0 －17.7 Wet cyanobacteria Lake Vanda X4 －15.6 －17.4 Dried cyanobacteria Lake Vanda X5 －15.6 －15.9 Dried cyanobacteria Lake Canopus C2 －18.5 －4.7 Dried cyanobacteria Lake Canopus C3 －9.6 －3.1 Wet cyanobacteria Lake Canopus C4 －11.7 －11.8 Wet cyanobacteria Lake Canopus C8 －11.0 －1.9 Wet cyanobacteria Lake Canopus C9 －14.7 －3.0 Dried cyanobacteria Don Juan Pond D1 －1.8 －19.7 Dried cyanobacteria Don Juan Pond D2 －13.0 －19.1 Dried cyanobacteria Don Juan Pond D3 －15.1 －20.2 Dried cyanobacteria Unnamed pond in the Labyrinth L11-A －4.5 －50.0 Dried one Unnamed pond in the Labyrinth L11 Black －5.6 －46.7 Wet cyanobacteria under the ice at 5 cm depth Unnamed pond in the Labyrinth L12 －6.5 －51.3 Wet cyanobacteria Unnamed pond in the Labyrinth L13 －7.7 －52.6 Wet cyanobacteria Unnamed pond in the Labyrinth L13 －6.5 －62.5 Wet cyanobacteria under the ice at 3 cm depth Unnamed pond in the Labyrinth L13 10 cm black －17.1 －9.3 Wet cyanobacteria under ice Unnamed pond in the Labyrinth L14 －13.3 －63.4 Wet cyanobacteria Unnamed pond in the Labyrinth L20 lichen A －25.2 －17.3 lichen Unnamed pond in the Labyrinth L20 lichen B －25.2 －20.4 Degraded lichen Taylor Valley East L. Bonney B1 －9.3 －76.5 Wet cyanobacteria near river mouth Melt-stream from Hughes Glacer B2 －12.7 －10.0 Wet cyanobacteria Melt-stream from Hughes Glacer B3 －12.7 －9.4 Wet cyanobacteria