Vol. 56: 65–79, 2009 AQUATIC MICROBIAL ECOLOGY Published online July 9, 2009 doi: 10.3354/ame01324 Aquat Microb Ecol OPPEENN ACCCCEESSSS Role of mesoscale cyclonic eddies in the distribution and activity of Archaea and Bacteria in the South China Sea Yao Zhang1, Eva Sintes2, Jianing Chen1, Yong Zhang1, Minhan Dai1, Nianzhi Jiao1, Gerhard J. Herndl2,3,* 1State Key Laboratory for Marine Environmental Sciences, Xiamen University, 422 Siming South Road, Xiamen 361005, China 2Department of Biological Oceanography, Royal Netherlands Institute for Sea Research (NIOZ), PO Box 59, 1790 AB Den Burg, The Netherlands 3University of Vienna, Ecology Center, Department of Marine Biology, Althanstr. 14, 1090 Vienna, Austria ABSTRACT: We examined the effect of 2 cold-core cyclonic eddies on the activity of the prokaryotic plankton community in the South China Sea. The abundance of bulk prokaryotes and major prokary- otic groups (Bacteria, marine Crenarchaeota Group I and marine Euryarchaeota Group II) and the number of cells taking up D- vs. L-aspartic acid (Asp) were determined using microautoradiography combined with catalyzed reporter deposition fluorescence in situ hybridization (MICRO-CARD- FISH). At all sites, the bulk D-Asp:L-Asp uptake ratio by the prokaryotic community increased with depth. Concurrently, the contribution of marine Crenarchaeota Group I to total prokaryotic abun- dance and total active cells also increased with depth, while an opposite pattern was observed for Bacteria. Marine EuryarchaeotaGroup II were generally more dominant in near-surface than in deep waters. Significant differences were observed between sites inside and outside the cyclonic eddies in terms of distribution and activity of prokaryotic communities and the concentration of fluorescent dis- solved organic matter (FDOM, an important refractory fraction of DOM). Generally, higher bulk D- Asp:L-Asp uptake ratios by the prokaryotic community and a greater crenarchaeotal contribution were found in the upper mesopelagic water column inside the cold-core eddies as compared to the outside sites. Taken together, the MICRO-CARD-FISH data and the pattern of FDOM indicate that the higher contribution of refractory DOM induced by upwelled water in the cyclonic eddy may have led to a more prominent role of Crenarchaeota in the organic carbon cycling in the mesopelagic realm of the cold-core eddy than outside the eddy in the South China Sea. KEY WORDS: Cyclonic eddy · Archaea · Bacteria · Enantiomeric amino acid uptake · DOM fluorescence Resale or republication not permitted without written consent of the publisher INTRODUCTION response of phyto- and microzooplankton (Allen et al. 1996, Benitez-Nelson et al. 2007, McGillicuddy et al. The presence of mesoscale cyclonic eddies can have 2007). Recently, Ewart et al. (2008) reported on the a significant impact on regional biogeochemical pro- bacterial biomass and production associated with cesses due to the introduction of new nutrients into the eddy-induced upwelling in the surface waters of the photic zone that leads to elevated primary production Sargasso Sea. However, little is known about the influ- and altered efficiency of the biological pump as com- ence of mesoscale cyclonic eddies that lead to an pared to the surrounding waters. Previous studies on upwelling of deep waters on the distribution and het- mesoscale cyclonic eddies have focused largely on erotrophic activity of Archaea vs. Bacteria in the their physical and nutrient dynamics and/or the mesopelagic water column. *Corresponding author. Email: [email protected] © Inter-Research 2009 · www.int-res.com 66 Aquat Microb Ecol 56: 65–79, 2009 Quantitative studies using fluorescence in situ scale cold-core eddies modified the commonly ob- hybridization (FISH) have revealed that in the ocean’s served vertical profile of prokaryotic community com- interior, planktonic Archaea(Crenarchaeotaand Eury- position and activity in the mesopelagic water column archaeota) might account for about one-third of all pro- of the SCS. karyotic cells in the global ocean (Karner et al. 2001), and that marine Crenarchaeota Group I are ubiqui- tously distributed in the deep ocean (Herndl et al. MATERIALS AND METHODS 2005, Kirchman et al. 2007, Varela et al. 2008). Also, studies have shown that, at least, some Crenarchaeota Sampling area. The SCS with its deep basin, is one are chemoautotrophic, utilizing inorganic carbon as a of the largest marginal seas in the tropical Pacific (Hu carbon source (Herndl et al. 2005, Wuchter et al. 2006) et al. 2000). Its basin-scale circulation is dominated by and oxidizing ammonia as an energy source (Könneke both the East Asian monsoon and the Kuroshio, which et al. 2005). Other studies have reported substantial is the subtropical western boundary current of the heterotrophy in Crenarchaeota as indicated by the North Pacific. Mesoscale eddies are a prominent fea- uptake of amino acids (Ouverney & Fuhrman 2000, ture of the SCS. Four to 6 mesoscale eddies are present Teira et al. 2006a). At present, a mechanistic under- at any given time over the deep basin of the SCS standing of thefactors controlling the distribution and (Hwang & Chen 2000, Wang et al. 2003). Driven by the extent of hetero- vs. autotrophic activity of the main prevailing southwesterly monsoonal winds in summer, archaeal group, the Crenarchaeota, is still largely lack- cold-core cyclonic eddies are formed in the western ing (Agogué et al. 2008). SCS (Hu et al. 2000). Among the members of the prokaryotic community, High-resolution surveys (Fig. 1B) were carried out to the marine CrenarchaeotaGroup I have been shown to localize eddies during the GOE-2 cruise on board the be mainly responsible for the uptake of D-antiomeric RV ‘Dongfanghong’ #2 on 14 Aug to 14 Sept 2007. For aspartic acid (D-Asp) in meso- and bathypelagic waters (Varela et al. 2008). The ratio of D:L -enantiomeric amino acids in dissolved organic matter (DOM) might serve as an indicator of its bioreactivity (Pérez et al. 2003), in addi- tion to the concentration of fluorescent dissolved organic matter (FDOM) (McKnight & Aiken 1998, Yamashita & Tanoue 2008). In this study, we deter- mined the distribution of the main prokaryotic groups (Bacteria, marine Crenarchaeota Group I and marine EuryarchaeotaGroup II) in the western South China Sea (SCS), which is an area characterized by prominent sea- sonal mesoscale eddies (Hwang & Chen 2000, Wang et al. 2003). In addi- tion, D- vs. L-aspartic acid uptake rates were determined to test for differences in DOM availability within and outside Fig. 1. (A) Map of the South China Sea showing the stations (Q) sampled for mi- the eddies, as proposed by Pérez et al. croautoradiography and catalyzed reporter deposition fluorescence in situhy- (2003) and Teira et al. (2006b) who bridization (CARD-FISH) (CE1 at 111.83°E, 14.25°N; CE2 at 111.03°E, 12.03°N; showed that high D-Asp:L-Asp uptake CEP at 113.00°E, 15.00°N; and SEATS at 115.96°E, 18.03°N). The water depths at these 4 sites are 2844, 2418, 2778 and 3839 m, respectively. The background ratios are indicative of a prokaryotic is a mosaic base map of real-time mesoscale sea-surface height altimetry remote community that is adapted to utilize sensed image with contours (http://argo.colorado.edu/~realtime/gsfc_global- old, more refractory DOM. By compar- real-time_ssh/) for 26 August 2007 (date sample from CE1 was taken) (the whole ing the distribution of the main image), 5 September 2007 (date sample from CE2 was taken) (108 to 112°E, 9 to prokaryotic groups, their uptake of D- 13°N) and 13 September 2007 (date sample from SEATS1 was taken) (114 to 120°E, 15 to 21°N). -- -- --: Boundaries of different images; CE1: cold-core vs. L-Asp and the vertical profiles of cyclonic eddy #1; CE2: cold-core cyclonic eddy #2; CEP: cold-core cyclonic eddy FDOM inside and outside the cyclonic periphery; SEATS: a time-series station in the South China Sea. (B) All survey eddy systems, we found that the meso- sites during the cruise (d) Zhang et al.: Archaeaand Bacteriain cyclonic eddies 67 this study, 4 stations were established. At each station, Uptake of D-Asp and L-Asp by the bulk prokaryotic a vertical profile with 8 depths from the surface to community. Triplicate 40 ml samples were incubated meso- and bathypelagic waters was sampled. Two sta- with D-[2,3-3H]-Asp (1 nM final concentration; 36 Ci tions were located inside cold-core cyclonic eddies mmol–1, Amersham) and L-[2,3-3H]-Asp (1 nM final (CE1 and CE2), one station was at the cold-core eddy concentration; 37 Ci mmol–1, Amersham) in the dark at periphery (CEP), and another was at the Southeast in situ temperature for 8 h (Teira et al. 2006a). Tripli- Asia Time-Series Study station (SEATS) (Fig. 1A). cate killed controls were fixed with 2% paraformalde- Hydrographic parameters.A CTD-General Oceanic hyde (Sigma) prior to substrate addition. Incubations rosette sampler with Go-Flo bottles (SBE 9/17 plus, were also terminated with 2% paraformaldehyde. SeaBird) was used to record temperature and salinity, Fixed samples were filtered as described above, rinsed and to collect water samples. Samples for inorganic twice with 0.2 µm filtered seawater and stored in scin- nutrients (nitrate + nitrite, phosphate, silicate) were fil- tillation vials at –20°C until analysis. Radioactivity was tered through 0.45 µm cellulose acetate filters and determined as described above for leucine incorpora- measured immediately onboard using a flow injection tion, and the DPM converted to D- and L-Asp uptake analyzer (Tri-223 autoanalyzer) and standard spec- rates. The analytical error of the estimates was <10%. trophotometric methods (Pai et al. 1990). Oxygen con- Abundance of Archaeaand Bacteriadetermined by centrations were determined onboard using the Win- CARD-FISH.Samples of 40 ml were immediately fixed kler method (Carpenter 1965). Apparent oxygen with freshly prepared paraformaldehyde (2% final utilization (AOU) and oxygen saturation (O S) were concentration) and stored at 4°C overnight prior to fil- 2 estimated based on in situO , temperature and salinity tration onto 0.2 µm pore size polycarbonate filters 2 (Garcia et al. 2006). Samples for chl a analysis were (Whatman). Filtered samples were stored at –20°C for collected on 0.7 µm pore-size GF/F filters (Whatman) later analysis by CARD-FISH. Picoplankton abun- and chl awas determined using a fluorometer (Turner dance was determined by DAPI staining, and Archaea Designs, Model 10). Chl a data were provided by the and Bacteria were enumerated by CARD-FISH. Fil- GOE (Group of Excellent) project (B.Q. Huang, Xia- ters were embedded in low-gelling-point agarose men University, China). and incubated either with lysozyme for the Bacteria To assess the evolution of the cold eddies, the sea probe mix (Eub338: 5’-GCT GCC TCC CGT AGG surface height anomaly was monitored through real- AGT-3’, Eub338II: 5’-GCA GCC ACC CGT AGG TGT- time mesoscale altimetry using images acquired from 3’ and Eub338III: 5’-GCT GCC ACC CGT AGG TGT - the Colorado Center for Astrodynamics Research 3’) and for the negative control probe (Non338: 5’-ACT (CCAR, USA). CCT ACG GGA GGC AGC-3’) or with proteinase-K Picoplankton production determined by 3H-leucine for the marine Crenarchaeota Group I probe mix incorporation.To determine heterotrophic picoplank- (Cren537: 5’-TGA CCA CTT GAG GTG CTG-3’ and ton production, triplicate 10 ml samples were incu- Cren554: 5’-TTA GGC CCA ATA ATC MTC CT-3’) bated with [3H]leucine (10 nM final concentration; 65 and for the marine Euryarchaeota Group II probe Ci mmol–1, Amersham) in the dark at in situtempera- (Eury806: 5’-CAC AGC GTT TAC ACC TAG-3’) (De- ture for 4 h (Herndl et al. 2005). Triplicate killed con- Corte et al. 2009). Filters were cut into 8 pieces and trols were fixed with 2% paraformaldehyde (Sigma) hybridized with horseradish peroxidase (HRP)-labeled prior to [3H]leucine addition. Incubations were termi- oligonucleotide probes and tyramide Alexa488 for sig- nated by adding 2% paraformaldehyde. The fixed nal amplification following the protocol described by samples were filtered (<0.3 atm) onto 0.2 µm pore size, Teira et al. (2004). white polycarbonate filters supported by 0.45 µm nitro- Enumeration of D-Asp and L-Asp active archaeal cellulose filters (Whatman), then rinsed 3× with 10 ml and bacterial cells by MICRO-CARD-FISH. Microau- of 5% ice-cold trichloroacetic acid, dried and placed in toradiography was applied on samples collected at sta- scintillation vials. Samples were stored at –20°C until tions CE1 and CEP. Triplicate samples of 40 ml were analysis in the laboratory. Filters were dissolved in incubated with D-[2,3-3H]-Asp (10 nM final concen- 1ml of ethyl acetate (Sigma) for 10 min, 8 ml of scintil- tration; 36 Ci mmol–1, Amersham) and L-[2,3-3H]-Asp lation cocktail (UltimaGold) was added, and counting (10 nM final concentration; 37 Ci mmol–1, Amersham) in a liquid scintillation counter (1220 QuantulusTM, in the dark at in situ temperature for 8 h (Teira et al. Wallac) was done after 12 h. The disintegrations per 2006a). Killed controls were fixed with 2% para- minute (DPM) of the paraformaldehyde-fixed blanks formaldehyde (Sigma) prior to substrate addition. Incu- were subtracted from the mean DPM of the respective bations were terminated with 2% paraformaldehyde samples, and the resulting DPM converted into leucine and stored in the dark at 4°C for 12 h. Thereafter, sam- incorporation rates. The analytical error of the esti- ples were filtered (<0.3 atm) onto 0.2 µm pore size, mates was <10%. white polycarbonate membrane filters supported by 68 Aquat Microb Ecol 56: 65–79, 2009 0.45 µm nitrocellulose membranes (Whatman), rinsed nine sulfate units (QSU) where 1 QSU is equivalent to twice with Milli-Q water and stored at –20°C until the fluorescence of 1 µg l–1quinine sulfate solution. For analysis. After the CARD-FISH procedure, hybridized comparison, series of duplicate samples were also filter sections were prepared for microautoradiography measured in a 1 ×1 cm quartz cell with a spectrofluo- (Teira et al. 2004). A glass slide was dipped into a rometer (Cary Eclipse, Varian) at 350 nm excitation molten (43°C) solution of NTB-2 emulsion (Kodak) and 450 nm emission using 20 nm bandwidths (Ferrari diluted to 2 parts of emulsion and 1 part of Milli-Q & Dowell 1998). Results obtained with the 2 instru- water. The filters with the cells were placed in contact ments were not significantly different (y = 0.9207x + with the molten emulsion. The glass slides were then 0.1066, r2= 0.966; p < 0.01). placed on an ice-cold aluminum block for 10 min to Statistical analysis. Since normality of distribution of solidify the emulsion before being transferred to light- the individual data sets was not always met, we used tight boxes containing a drying agent for autoradi- the non-parametric Wilcoxon and Friedman tests for ographic exposure at 4°C for 7 d. The emulsion was comparing 2 and >2 related variables, respectively. developed following Kodak’s specifications. Before Model II Reduced Major Axis (RMA) regression, expo- complete drying, filter sections were removed and nential growth and decay models (Sigmaplot) were cells were counterstained with a DAPI mix: 5.5 parts of used to determine the relationships between variables. Citifluor (Citifluor), 1 part of Vectashield (Vector Labo- ratories) and 0.5 parts of phosphate-buffered saline (PBS) with DAPI (final concentration of 5 µg ml–1). RESULTS Microscopy and image analysis. The slides were assayed by microscopy using images acquired by flu- Hydrographic characteristics orescence and transmitted light, and a digital camera (AxioCam MRc5) mounted on an epifluorescence Two well-developed cold-core cyclonic eddies (CE1 microscope (Zeiss Axioplan 2) equipped with a 100 W and CE2) were identified and their development was mercury lamp and a tungsten lamp. Cells in 10 to 30 tracked via satellite altimetry (Fig. 1). Shipboard microscopic fields were examined per sample. Three Acoustic Doppler Current Profiler (ADCP) data docu- images were acquired for each field. DAPI and mented the counterclockwise currents in CE1 and Alexa488 images were acquired using appropriate fil- CE2, which are associated with their negative sea- ter sets with optimized manual exposure time of 400 level anomaly. Their altimetric history suggested to 1900 ms. Negative control counts (hybridization intensification of CE1 between 14 and 30 Aug, and of with HRP-Non338) were always <1% of DAPI-stained CE2 between 26 Aug and 12 Sept. Observations were cells. Images of silver grains were obtained using conducted during the intensification phases as samples transmitted light with automatic exposure. In the were taken at CE1 on 26 Aug, at CEP on 28 Aug, at killed controls, <0.5% of the total DAPI-stained cells CE2 on 5 Sept and at SEATS on 13 Sept (Fig. 1A). were associated with silver grain halos. More than The cores of CE1 and CE2 were located at 500 DAPI- stained cells were counted per sample. ~112.5°E, 13.5°N and 111°E, 12°N, respectively. A The 3 images were overlaid to obtain composite comparison of the vertical profiles of physical and images to identify DAPI-stained cells, probe-positive chemical parameters in the 4 stations show that the cells and cells with silver grains. For each micro- shallow waters (upper 200 m) of CE1 and CE2 exhib- scopic field, 4 categories were differentiated: (1) total ited lower temperature, higher salinity, higher nutri- DAPI-stained cells, (2) cells stained with a specific ent concentrations (M. Dai et al. unpubl.), lower dis- probe, (3) DAPI-stained cells with associated silver solved oxygen, and higher chl aconcentration (Fig. 2) grain halos, and (4) cells labeled with a specific fluo- and primary production (data not shown) than Stns rescent probe and silver grain halos associated with CEP and SEATS. CEP was located at the cold-core individual cells. The analytical error of the estimates eddy periphery and hence, was potentially influenced was <10%. by the mesoscale features. SEATS was located far out- Measurement of FDOM. The FDOM was measured side the 2 eddy systems (Fig. 1). The hydrographic as a proxy for biorefractory DOM in subsurface waters data from SEATS presented here (Fig. 2) are within (Yamashita & Tanoue 2008). Seawater was filtered the range of historical data (www.ncor.ntu.edu.tw/ through pre-cleaned 0.22 µm pore size polycarbonate SEATS/SEATS_Eng.htm), indicating that SEATS is a membrane filters (Millipore). Fluorescence was mea- representative outside-the-eddy station. The surface sured onboard using the Turner Designs (Model 10- waters of CE2 were influenced by the Mekong River AU) fluorometer equipped with an FDOM optical kit at plume and had relatively low salinity. At all 4 sites, 310 to 390 nm excitation and 400 to 600 nm emission. oxygen minimum layers were detectable at ~800 m The fluorescence of the samples is expressed in qui- depth. Zhang et al.: Archaeaand Bacteriain cyclonic eddies 69 CE1 CE2 CEP SEATS Oxygen (µmol kg–1) Oxygen (µmol kg–1) Oxygen (µmol kg–1) Oxygen (µmol kg–1) 50 100 150 200 250 50 100 150 200 250 50 100 150 200 250 50 100 150 200 250 Temperature (°C) Temperature (°C) Temperature (°C) Temperature (°C) 0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30 0 0 0 0 50 50 50 50 100 100 100 100 m) 150 150 150 150 pth ( 250000 Temp. 250000 250000 250000 e D 1000 Salinity 1000 1000 1000 1500 1500 1500 1500 DO 2000 2000 2000 2000 32.5 33.5 34.5 32.5 33.5 34.5 32.5 33.5 34.5 32.5 33.5 34.5 Salinity Salinity Salinity Salinity Silicate (µmol l–1) Silicate (µmol l–1) Silicate (µmol l–1) Silicate (µmol l–1) 0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30 Nitrate + Nitrite (µmol l–1) Nitrate + Nitrite (µmol l–1) Nitrate + Nitrite (µmol l–1) Nitrate + Nitrite (µmol l–1) 0 10 20 30 0 10 20 30 0 10 20 30 0 10 20 30 0 0 0 0 50 50 50 50 m) th ( 100 NChl a 100 100 100 p e P D Si 150 150 150 150 200 200 200 200 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 0 1 2 3 4 5 Phosphate (µmol l–1) Phosphate (µmol l–1) Phosphate (µmol l–1) Phosphate (µmol l–1) 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 0.0 0.1 0.2 0.3 0.4 0.5 Chl a (µg l–1) Chl a (µg l–1) Chl a (µg l–1) Chl a (µg l–1) Fig. 2. Physical and chemical characteristics of the different sites sampled in the South China Sea in summer 2007. Descriptions of stations are as in Fig. 1. The chl a data from samples taken at 6:30 to 7:00 h are plotted. Nutrient concentration values of 0 refer to values below the limit of detection of standard spectrophotometric methods Prokaryotic abundance and activity inside and distribution of chl aover the euphotic layer between the outside the cyclonic eddies inside and outside stations coincided with the distribu- tion pattern of prokaryotic abundance at Stns CE1 and Prokaryotic abundance decreased exponentially CE2 vs. CEP and SEATS. While differences in prokary- with depth, ranging between 4.3 ×105 and 7.5 ×105cells otic abundance among the 4 stations were apparent in ml–1at 50 m depth and from 0.6 ×105 to 1.2 ×105cells surface waters, no significant differences were found in ml–1at 1500 to 2000 m depth. Differences in prokaryotic the meso- and bathypelagic layers. abundance in the shallow waters (≤100 m) of the 4 sites Leucine incorporation, as a measure of prokaryotic were apparent. The high prokaryotic abundance at 50 activity, decreased with depth by 2 orders of magni- and 100 m depth in CEP and SEATS (Fig. 3A), respec- tude (Table 1). Prokaryotic activity at 50 and 100 m tively, coincided with high chl aconcentrations at these depth was higher at CEP and SEATS than in the eddy depths. At CE1 and CE2, upwelled nutrient-rich water centers. In the deep layers of the water column (800 to supported a higher phytoplankton biomass in the photic 2000 m depth), prokaryotic activity ranged from 0.43 to zone, and the chlorophyll maximum layer was shallower 1.02 and from 0.18 to 0.45 pmol l–1h–1inside (CE1 and (Fig. 2) than outside the eddies. These differences in the CE2) and outside (CEP and SEATS) the eddy centers, 70 Aquat Microb Ecol 56: 65–79, 2009 CE1 CE2 CEP SEATS Bacterial abundance (x104 cells ml–1) A 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 0 10 20 30 40 50 50 50 50 100 100 100 100 m) 200 200 200 200 h ( 300 300 300 300 pt 450 450 450 450 e D 800 800 800 800 1000 EUB 1000 1000 1000 2000 DAPI 1500 1500 2000 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 0 2 4 6 8 10 DAPI counts (x105 cells ml–1) Archaeal abundance (x104 cells ml–1) B 0 2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 12 0 2 4 6 8 10 50 50 50 50 100 100 100 100 m) 200 200 200 200 h ( 300 300 300 300 pt 450 450 450 450 e D 800 800 800 800 1000 CREN 1000 1000 1000 2000 EURY 1500 1500 2000 % of prokaryotic cells C 0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50 0 10 20 30 40 50 50 CREN 50 50 50 EURY 100 EUB 100 100 100 m) 200 200 200 200 h ( 300 300 300 300 pt 450 450 450 450 e D 800 800 800 800 1000 1000 1000 1000 2000 1500 1500 2000 Fig. 3.(A,B) Abundances (mean ±SDs) of total prokaryotes (DAPI-stained cells), Bacteria(EUB), marine CrenarchaeotaGroup I (CREN) and marine Euryarchaeota Group II (EURY), and (C) their contributions to total prokaryotic abundance inside and outside the cyclonic cold eddies in the South China Sea. Descriptions of stations are as in Fig. 1 respectively, while in the intermediate layers (200 to teria, marine CrenarchaeotaGroup I and marine Eury- 450 m depth), prokaryotic activity was not significantly archaeotaGroup II. The numbers of DAPI-stained cells different among the 4 sites. before and after CARD-FISH processing were not sig- nificantly different (data not shown), indicating that cell loss during CARD-FISH processing was negligible. Prokaryotic community composition inside and Generally, total prokaryotic and bacterial abundance outside the cyclonic eddies decreased similarly with depth at all 4 sites (Fig. 3A). Also, the abundance of marine EuryarchaeotaGroup II Prokaryotic community structure was examined by decreased with depth although in a less pronounced CARD-FISH. On average, 63 ±5% of the total number way than Bacteria, while marine CrenarchaeotaGroup of prokaryotes (DAPI-stained cells) was detected by I did not exhibit a depth-related trend (Fig. 3B). There CARD-FISH with HRP oligonucleotide probes for Bac- were, however, remarkable detectable differences in Zhang et al.: Archaeaand Bacteriain cyclonic eddies 71 Table 1. Leucine incorporation rates (Leu) and L-antiomeric aspartic acid (L-Asp) uptake rates (pmol l–1 h–1). CE1 and CE2: cold-core cyclonic eddies, CEP: cold-core eddy periphery, and SEATS: Southeast Asia Time-Series Study station Depth (m) CE1 CE2 CEP SEATS Leu 50 17.08 13.95 18.62 14.95 100 3.54 3.16 4.68 4.65 200 – 450a 1.07 ±0.12 1.29 ±0.47 0.86 ±0.18 1.09 ±0.24 800 – 2000a 0.76 ±0.18 0.80 ±0.32 0.310 ±0.098 0.34 ±0.14 L-Asp 50 1.76 1.03 2.41 1.41 100 0.41 0.34 0.71 0.71 200 – 450a 0.096 ±0.013 0.096 ±0.021 0.090 ±0.028 0.097 ±0.054 800 – 2000a 0.026 ±0.011 0.0480 ±0.0097 0.019 ±0.0045 0.025 ±0.0049 aMean values and SDs were calculated from data at 3 depths the vertical patterns of the relative abundances of the 3 depth at CE1 and at 1500 m depth at CE2 (Fig. 4B). main prokaryotic groups (Fig. 3C). While the contribu- Corresponding to the higher bulk D-Asp uptake rates tion of Bacteria to total picoplankton abundance in the 2 eddies (Friedman, p < 0.0001), the D-Asp:L- decreased with depth, the relative abundance of Asp uptake ratios were significantly higher inside than marine CrenarchaeotaGroup I increased, contributing outside the cyclonic eddies (Friedman, p < 0.0001), between 6 and 10% to total picoplankton abundance especially in intermediate waters (up to ~5× at 100 to at 50 m and from 24 to 32% at 1500 to 2000 m depth. 450 m depth). The D-Asp:L-Asp uptake ratio de- The variation in relative abundance of marine Eury- creased exponentially with increasing leucine incor- archaeota Group II with depth was small, except in poration rates (Fig. 4C). SEATS, where a clear decreasing trend was observed down to 450 m depth. Marine Crenarchaeota Group I were more abundant than marine Euryarchaeota Distribution of D-Asp and L-Asp uptake among Group II in meso- and bathypelagic waters at all 4 sites Archaeaand Bacteria (below 200 m depth) (Wilcoxon, p < 0.05 for both abun- dance and percentage at each site). In shallow waters The distribution of Archaea (marine Crenarchaeota (≤100 m), however, Euryarchaeota were more abun- Group I and Euryarchaeota Group II) and Bacteria dant at CEP and SEATS (even more abundant than capable of taking up D-Asp vs. L-Asp was examined at Crenarchaeota) than at CE1 and CE2 (Fig. 3B). The CE1 and CEP by MICRO-CARD-FISH. The recovery percentage of total picoplankton identified as marine efficiencies of picoplankton cells that took up L-Asp CrenarchaeotaGroup I was significantly higher at CE1 and D-Asp were 93 ± 4 and 88 ± 6%, respectively, and CE2 than at CEP and SEATS throughout the water using the HRP oligonucleotide probes for Bacteria, column (Friedman, p < 0.0001), particularly in the shal- marine Crenarchaeota Group I and marine Eury- low and intermediate waters (up to ~2×at depths shal- archaeotaGroup II. The percentage of Bacteriataking lower than 800 m) (Fig. 3C). up L-Asp and D-Asp out of the total active picoplank- ton cells decreased with depth from ~60% at 50 m depth to ~40% at 1500 to 2000 m depth, while the con- D-Asp vs. L-Asp uptake by the bulk picoplankton tribution of marine Crenarchaeota Group I to total community picoplankton taking up L-Asp and D-Asp increased with depth from ~20% at 50 m depth to 30 and 40% in Generally, the uptake pattern of L-Asp was similar to the bathypelagic layers, respectively (Fig. 5). The con- leucine incorporation from surface to deep waters at tribution of Crenarchaeota to the total D-Asp+ cells the 4 sites (Table 1). D-Asp uptake rates decreased was significantly higher than to the total L-Asp+ cells from shallow waters to the oxygen minimum layer, but for both sites (Wilcoxon, p < 0.05 for both CE1 and increased again in bathypelagic waters (Fig. 4A). The CEP), in contrast to Euryarchaeota. Significantly D-Asp:L-Asp uptake ratio of the bulk picoplankton higher crenarchaeotal contributions to both total D- community increased with depth at all study sites, Asp and L-Asp+ cells were observed at CE1 than at ranging between 0.017 and 0.082 at 50 m depth and CEP (Wilcoxon, p < 0.05 for both D-Asp and L-Asp), from 0.63 to 1.41 at 1500 to 2000 m depth. D-Asp:L-Asp while the contribution of marine EuryarchaeotaGroup uptake ratios >1 were measured at 1000 to 2000 m II to L-Asp+ cells was significantly higher at CEP than 72 Aquat Microb Ecol 56: 65–79, 2009 CE1 CE2 CEP SEATS D-Asp uptake rate (pmol l–1 h–1) A 0.00 0.02 0.04 0.06 0.08 0.10 0.00 0.02 0.04 0.06 0.08 0.10 0.00 0.02 0.04 0.06 0.08 0.10 0.00 0.02 0.04 0.06 0.08 0.10 50 50 50 50 100 100 100 100 m) 200 200 200 200 h ( 300 300 300 300 pt 450 450 450 450 e D 800 800 800 800 1000 1000 1000 1000 2000 1500 1500 2000 D-Asp : L-Asp uptake ratio B 0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 50 50 50 50 100 100 100 100 m) 200 200 200 200 h ( 300 300 300 300 pt 450 450 450 450 e D 800 800 800 800 1000 1000 1000 1000 2000 1500 1500 2000 C 10 10 10 10 o rati 1 1 1 1 p s A - 0.1 0.1 0.1 0.1 L p : -As 0.01 r2 = 0.82 0.01 r2 = 0.83 0.01 r2 = 0.97 0.01 r2 = 0.92 D 0.001 0.001 0.001 0.001 0.1 1 10 100 0.1 1 10 100 0.1 1 10 100 0.1 1 10 100 Leu uptake rate (pmol l–1 h–1) Fig. 4. (A) D-Asp (aspartic acid) uptake rates (pmol l–1h–1), (B) D-Asp:L-Asp uptake ratios, and (C) correlation between D-Asp:L- Asp uptake ratios and leucine (leu) incorporation rates of the bulk prokaryotic community inside and outside the cyclonic cold ed- dies in the South China Sea. The fitting function used between the D-Asp:L-Asp uptake ratio and leucine incorporation is y= aebx. Descriptions of stations are as in Fig. 1. Error bars indicate ±SD at CE1 (Wilcoxon, p < 0.05). No significant differences Group II decreased with depth, while the D-Asp active in euryarchaeotal contribution to D-Asp+ cells be- fractions did not exhibit depth-related trends (<15%). Of tween the 2 sites (Wilcoxon, p > 0.1) were found. the total picoplankton cells taking up L-Asp, marine Cre- The fractions of D-Asp and L-Asp+ Bacteria, Crenar- narchaeotaGroup I and marine EuryarchaeotaGroup II chaeotaand Euryarchaeotaof the total number of Bacte- accounted for ~20% in the shallow waters of CE1 and ria, Crenarchaeotaand Euryarchaeotawere also deter- CEP (Fig. 5); however, ~50% of all the Crenarchaeota mined (Table 2). Generally, the L-Asp active fractions and Euryarchaeotatook up L-Asp (Table 2). In the ba- were significantly higher than the D-Asp active fractions thypelagic waters of both sites, only <10% of the total of the total bacterial, crenarchaeal and euryarchaeal crenarchaeotal and euryarchaeotal cells took up L-Asp. cells at the 2 sites (Wilcoxon, p < 0.05 for the 3 groups in More than 45% of Bacteria were L-Asp+ in shallow both CE1 and CEP). The L-Asp active fractions of Bacte- waters, compared to <10% of Bacteriabeing L-Asp+ in ria, marine CrenarchaeotaGroup I and Euryarchaeota deep waters. The ratios of D-Asp:L-Asp+ cells increased Zhang et al.: Archaeaand Bacteriain cyclonic eddies 73 CE1 CEP CE1 CEP % of L-Asp+ cells Ratio of D-Asp+:L-Asp+ cells 0 20 40 60 80 100 0 20 40 60 80 100 0.0 0.5 1.0 1.5 2.0 0.0 0.5 1.0 1.5 2.0 50 50 50 50 100 100 100 100 CREN CREN m) 200 200 m) 200 EURY 200 EURY h ( 300 300 h ( 300 EUB 300 EUB ept 450 450 ept 450 450 D D 800 800 800 800 1000 CREN 1000 1000 1000 EURY 2000 EUB 1500 2000 1500 % of D-Asp+ cells Fig. 6. Ratio of D-Asp:L-Asp (aspartic acid) active cells in Bac- 0 20 40 60 80 100 0 20 40 60 80 100 teria (EUB), marine Crenarchaeota Group I (CREN) and marine Euryarchaeota Group II (EURY). CE1: cold-core cy- 50 50 clonic eddy #1; CEP: cold-core cyclonic eddy periphery. Val- ues were calculated by dividing the means of D-Asp+ frac- 100 100 tions by the means of L-Asp+ fractions of major prokaryotic m) 200 200 groups (no error bars). The value of 100% is the total crenar- h ( 300 300 chaeotal, euryarchaeotal or bacterial probe positive DAPI- pt 450 450 stained cells e D 800 800 and SEATS (Friedman, p < 0.01), especially in the shal- 1000 1000 low and intermediate waters (≤500 m) (Fig. 7). The 2000 1500 ratio of humic-like FDOM to bulk DOC concentrations (data not shown) was higher in the shallow/intermedi- ate waters (25 to 450 m depth) in the eddies (range: 9.8 Fig. 5. Contributions (%, mean ± SD) of Bacteria (EUB), to 18 QSU (µM DOC)–1) than outside the eddies (range: marine Crenarchaeota Group I (CREN) and marine Eury- 2.3 to 15 QSU (µM DOC)–1) (Wilcoxon, p < 0.05). For archaeotaGroup II (EURY) tothe total D- and L-Asp+ cells of bathypelagic waters, no significant difference in DOC- communities. CE1: cold-core cyclonic eddy #1; CEP: cold- core cyclonic eddy periphery. The value of 100% is the total normalized fluorescence was found between stations D- or L-Asp+ DAPI-stained cells. Asp: aspartic acid inside and outside the eddies. with depth at both CE1 and CEP Table 2. Fraction (%, mean of duplicate measurements) of D-Asp+ and L-Asp+ for Bacteria, marine Crenarchaeota Bacteria(EUB), Crenarchaeota(CREN) and Euryarchaeota(EURY) of the total number of Bacteria, Crenarchaeotaand Euryarchaeota Group I and Euryarchaeota Group II (Fig. 6). This ratio increased more with depth in marine CrenarchaeotaGroup I Depth (m) EUB CREN EURY D-Asp+ L-Asp+ D-Asp+ L-Asp+ D-Asp+ L-Asp+ than in Bacteria and marine Eury- archaeota Group II (Friedman, p < Stn CE1 0.0001 for both CE1 and CEP). How- 50 9.66 45.17 11.56 51.76 10.34 52.87 100 10.29 41.15 10.50 33.15 13.04 57.61 ever, no significant differences in the 200 8.77 32.46 7.93 21.34 13.21 58.49 ratio D-Asp:L-Asp+ cells of the 3 300 8.40 24.43 9.09 18.18 10.29 38.97 prokaryotic groups were found be- 450 7.25 16.31 9.28 14.43 10.26 31.03 800 6.11 12.78 8.75 10.42 8.16 23.47 tween the 2 sites (Wilcoxon, p > 0.5 for 1000 8.13 11.25 8.98 7.81 7.50 15.18 Crenarchaeotaand Euryarchaeota; p > 2000 11.45 9.04 12.33 6.67 8.37 8.84 0.05 for Bacteria). Stn CEP 50 14.49 88.31 11.22 52.79 8.20 31.88 100 8.98 37.89 7.69 28.85 5.41 22.39 200 10.63 42.95 14.06 40.63 9.59 40.75 FDOM distribution inside and 300 7.33 22.33 10.50 22.00 9.20 32.76 outside the cyclonic eddies 450 10.30 25.32 14.00 21.00 11.11 36.11 800 9.04 19.59 10.90 12.82 8.24 24.73 1000 6.28 8.70 7.93 6.61 6.90 15.52 Humic-type FDOM was generally 1500 10.22 8.06 11.62 5.99 9.09 10.69 higher at CE1 and CE2 than at CEP 74 Aquat Microb Ecol 56: 65–79, 2009 Humic-type fluorescence of DOM (QSU) 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 0.2 0.4 0.6 0.8 1.0 5 5 5 5 35 20 25 25 50 50 50 50 m) 75 75 70 75 h ( 100 100 100 100 pt e 200 150 200 200 D 500 450 1000 1000 CE1 CE2 CEP SEATS 1500 1300 Fig. 7.Vertical distribution of fluorescent dissolved organic matter (FDOM). Descriptions of stations are as in Fig. 1. Only samples from the upper ocean shallower than 200 m were taken for FDOM fluorescence measurement at CE2 and SEATS. Error bars indicate ±SD.QSU: quinine sulfate unit DISCUSSION et al. 2008). Although the multiple probes targeted Bacteria and Crenarchaeota, the recovery efficiency The SCS is a semi-enclosed marginal sea with a was well below 100% of DAPI-stained cells. The deep basin in the tropical–subtropical western North explanation might be that some cells did not contain a Pacific (Fig. 1). It is characterized by oligotrophic con- sufficient number of ribosomes, although the signal ditions, a shallow mixed layer and nutricline, low pri- amplification procedure was used in this study. A cer- mary production (23 to 42 mmol C m–2 d–1) (Liu et al. tain fraction of DAPI-stainable cells might also be dor- 2002) and low export production (1 to 3.4 mmol C m–2 mant or dead, and hence did not contain any ribosome d–1) (Chen et al. 1999). Driven by the prevailing mon- as shown previously (Heissenberger et al. 1996). soonal wind stress, mesoscale eddies are active in, and Our results revealed a steady increase in the contri- characterize the hydrography of, the SCS (Hwang & bution of marine Crenarchaeota Group I to the total Chen 2000, Wang et al. 2003). picoplankton abundance from shallow layers down to During the GOE-2 cruise, 2 well-developed cold- bathypelagic waters in the SCS (Fig. 3C), confirming core cyclonic eddies were tracked. Apparently, the dif- the previously reported general distribution pattern of ferences in water mass movement inside and outside marine CrenarchaeotaGroup I with depth (DeLong et the cyclonic eddies led to distinct temperature, salinity, al. 1999, Karner et al. 2001, Herndl et al. 2005, Teira et nutrient and dissolved oxygen characteristics (Fig. 2), al. 2006a, Varela et al. 2008); a contrasting pattern was which consequently modified picoplankton commu- found for Bacteria, which is similar to previous studies nity composition and activity. in the Pacific (DeLong et al. 1999, Karner et al. 2001). There is growing evidence that Crenarchaeota are particularly associated with low-oxygen environments Picoplankton community composition inside and such as upwelling areas (Coolen et al. 2007, Lam et al. outside cyclonic eddies 2007). In the present study, the contribution of marine Crenarchaeota Group I to total picoplankton abun- In this study, the recovery efficiency (sum of Bacte- dance throughout the water column was significantly ria, marine Crenarchaeota Group I and marine Eury- higher inside (up to 31% of DAPI-stained cells) than archaeota Group II as a percentage of DAPI-stained outside the cyclonic eddies (Friedman, p < 0.0001), cells) ranged between 53 and 76%, which is compara- especially in shallow and intermediate waters (upper ble with those reported for North Atlantic meso- and 800 m, up to ~2×) (Fig. 3C), coinciding with lower oxy- bathypelagic waters (Teira et al. 2006a, Varela et al. gen concentrations inside than outside the cyclonic 2008). There is one methodological difference between eddies (Fig. 2). Concurrently, significantly higher con- this study and the previous studies, as we used the centrations of nitrous oxide were found inside than oligonucleotide probe mix of Cren537 and Cren554 outside the eddies, especially in intermediate waters (De-Corte et al. 2009) to enumerate Crenarchaeota, shallower than 450 m depth (H. Lin & M. Dai unpubl.). whereas the previous studies used only Cren537. This Also, archaeal amoAgene copy numbers were at least might have resulted in a higher percentage of DAPI- one order of magnitude higher than betaproteobacter- stained cells identified as Crenarchaeotain the present ial amoA gene copy numbers (A. Hu & N. Jiao un- study than reported for the subtropical Atlantic (Varela publ.). This suggests that Crenarchaeotamight play a