Active mesopelagic prokaryotes support high respiration in the subtropical northeast Atlantic Ocean

GEOPHYSICAL RESEARCH LETTERS, VOL. 32, L03608, doi:l0.1029/2004GL021863, 2005 Active mesopelagic prokaryotes support high respiration in the subtropical northeast Atlantic Ocean Javier Arístegui Departamento de Biología. Universidad de las Palmas de Gran Canaria. Las Palmas de Gran Canaria. Srain Carlos M. Duarte Instituto Mediterráneo de Estudios Avanzadus (lMEDEA). Consejo Superior de Investigaciones Científicas U11Ivcrsidacl de las Islas Baleares (CSIC-UiB), Esporles. Srain Josep M. Gaso! and Laura A!onso-Sáez Departament de Biologia Marina i Oceanografía. Institut de Ciéncies del Mar . Centre Mediteranei d'lnvestigacions Marines i Ambientals (lCM-CM1MA), Consejo Superior de Investigaciones Científicas, Barcelona, Spain Rcccivcd 28 Octobcr 2004; rcvised 23 Dcccmber 2004; acccptcd (, January 2005; publishcd II FcbrualY 200S. [1] Here we provide evidence, based on prokaryote metabolic proxies and dircct estimates 01' oxygen consumption, thal Ihe mesopelagic prokaryote assemblage in the subtropical Northeast Atlantic is an active one. It supports a high respiration (0.22 • 0.05 pmol O2 r 1 di, corresponding to 68 ± 8 mmol CO 2 m 2 d' 1), comparable to that 01' the epipelagic zone during (he same period (6497 mmol e m 2 dI). Our findings suggest that mesopelagic prokaryotes in the NE subtropical Ocean, as well as in other eastern boundal)' regions, are important carbon sinks for organic malter advecled from lhe highly productive coastal syslems, and would playa key role in the global carbon cycle 01' the oceans. Citation: AlÍstegui, J., e. M. Duarte, J. M. Gasol, and L. Alonso-Sáez (2005), Active Illcsopclagic prokaryotes support high respiration in the subtropical northeast Atlanlic Ocean, Geophvs. Res. Lett., 32. L03608. doi:l 0.1 029/2004GL021863. 1. lntroduction [2] The mesopelagic (2001000 m) open occan has been suggestcd as a major site for biological oxygen consumption and CO2 production [del GiOlgio and Duarte, 2002]. Yet, oxygen consumption in the rnesopelagic zone has only been examined directly in a few regions, thereby precluding veritication of the presumed role 01' its biota on the carbon budget 01' the ocean. The pioneering work 01' Jenkins in the Sargasso Sea [Jenkins, 1977] led for the fírst time to lhe estimation 01' oxygen consumption in the mesopclagic zone of the ocean, ÍÍ'Ol11 changes in (he apparen! oxygen utilization (AOU) and lhe use of tracers to ealeulate the apparcnt age 01' a water mass. His results were surprising in tha( he showcd that oxygen consumptiol1 in the mesopclagic zone of ¡he Sargasso Sea was considerably higher than the estimates of new productiol1 in the same region [Jenkins, 1982]. Unfortunately, this biogeochemical approach is only useful for the fCw oceanic regions where mixing below the surface thennocline is assumed to be unimportant, and consequently the age of the water mass can be calculated with relative eonfidence. This handicap has impeded the Copyright 200S by Ihe American Gcophysical Unioll. 0094-8276!05/2004GL021863$05.00 derivation oí' global estimates 01' respiration in the dark ocean through AOU/traccrs. [3] The first assessment 01' global respiration in the dark ocean was recently produced by Arístegui et al. [2003]. These authors compiled available estimates 01' enzymatic ETS (clectron transport system) rcspiratOl)' activity in the dark ocean, and converted them to actual respiration rates (R), by using a constant RlETS ratio ofO.09. The ratio was derived fi'om empirical R/ETS relationships detemlined in vi/ro from monospeeific cultures oi' bacteria at their senescent phase [Chrislensen el al., 1980], suggested to be representative of the dark ocean [Packard el al., 1988]. The calculated average global R in the dark ocean (5 mol C m - 2 yr - 1) was in close agreel1lent with the oxygen utilization rates inferred frol1l large-scalc l:racer balances by Jenkins for the Sargasso Sea, building confidence 10 the R/ETS calculations. However, Arísteglli et al. [2003] cautioned about the uncertainty in the general use 01' a constant R/ETS ratio derived from bacteria cultures in a low-activity physiological state. Since, prokaJ)'otic organisms in the deep ocean could maintain a rather high activity [e.g., Cllo alld AzallJ, 1988], the ETS to R conversion factor would need to be validated in each region through concomitant oxygen consumption and enzymatic l1leasurements 01' natural mesopelagic coml1lunities. [4] Here, we have addressed this challenge by combining direct measurel1lents of in vitro oxygen consumption at selected depths and detailed vertical profíles 01' ETS de terminations, in the mesopelagic zone of the subtropical NOliheast Atlantie Ocean. Additionally, we examined differen! metabolic proxies í'or the degree 01' activity 01' (he prokaryote asscmblages in the mesopelagic zone, and compared them with lhe activity of surface-water assemblages. We aimed to test the hypothesis that eastern boundary regions, which are supposed to reccive high lateral inputs oí' organic matter from coastal upwelling ecosystems, would maintain highly active mesopclagic microbial assemblages, supporting a high respiratory activity. 2. Methods 2.l. Data Collcction [5] The study was conducted along two zonal sections (21°N and 26°N) extending from the NW African shelf to the open Atlantic Ocean at 26°W, during May-June 2003. L03608 1 01' 4 L03608 ARÍSTEGUI ET AL MESOPELAGIC PROKARYOTES IN THE NE ATLANTIC L03608 Table 1. Prokaryotc Abundance and Activiry in the Epipclagic and Mcsopclagic Zones E~i[>clagic Zone Meso[>cla~ic lone N Prokaryote abllndance (10' Cells 1 1) 23 High-NA-content prokaryotcs (% of total) 23 leucine/Thymidinc-bascd P (nmols C lid ') 27 Prokaryote P lB (d 1) 27 erc + prokaryotc ablllldancc (lO' Cells 1-1) 23 erc + Eroka!10tes (% of total) 23 A total of 10, roughly equidistant, stations (5 in each section) were occupied. Samples for detennination of prokaryote abundance and their physiological state were collected each 100 m down to 1000 m. Samples for respiration experiments were obtained from 600 m (10 ± O.I°C) and 1,000 m (6.7 ± 0.1 OC) depth. and immediately taken to chambers set at the in situ temperature (±0.2°C) conditions. 2.2. Prokaryote Respiration (R) [6] Two approaches were used to estimate R: (1) Water samples drawn into carboys were siphoned into 5 replicate "time-zero" and "dark" (incubated for 2-4 days) 125 mlBOD bottles. A time-series experiment was carried out at the most oceanic station; with 2 extra sets of "dark" replicate bottles added at intennediate times. R was estimated [rom the difference in oxygen concentration between the "zero" and "dark" bottles following incubation. Dissolved oxygen was measured by the Winkler technique, using an automated precise titration system, with colorimetric end-point detection. The mean standard error (SE) among replicated bottles was 0.08 ~L1nol O2 r l. (2) Because R might often he beyond the detection Iimit for the procedure described, R was also assessed by pre-concentrating (on-average, a factor of 2.2 fold) the prokaryote assemblage by inverse filtration under gentle pressure, using a temperature-controlled eell concentration chamber fitted with 76 mm, 0.1 flm polycarbonate filters. The oxygen concentration was calculated spectrophotometrically, with a mean SE among replicated bottles of 0.43 pmol O2 11 and a mean SE for replicated l11easurements from the sal11e bottle of 0.19 fImol O2 1-1. R was calculated as aboye. Additionally, we monitored changes in abundanee and metabolism of the microbial assemblage during the incubations, allowing the backscaling of R to that corresponding to the assemblage prior to incubation (see results for details). 2.3. ETS Measurements [7] Seawater (10 to 20 1), collected every 50 m from 200 m to 1000 m depth, was tiltered through 47 mm Whatman GF/F filters, at a low vacuum pressure «0.3 atm). The filters were stored in liquid nitrogen until being assayed in lhe laboratory. ETS determinations were carried out as described by Arístegui and Montero [1995]. An incubation time of 20 min at 18°C was used. ETS activities measured at 18°C were converted to activities at in situ temperatures by using the Arrhenius equation. A calculated activation energy o[ 16 kcal mole I was used. 2.4. Prokaryote Abundance and Biomass [8] Cells were counted by flow cytometry, usmg a FACSealibur (Becton & Diekinson) instrument [Gasol and del Giorgio, 2000]. Sal11ples were fixed with 1 % of A\'Cra~c ± SE N Average:::: SE Ratio E:M 12A ± 0.3 31 1.R ± 0.2 6.9 37.8 ± 2.2 27 57.9 ± 2.0 0.7 391.6 ± 77.1 29 32.7 ± 6.5 12.4 OAO ± 0.12 29 0.13 ± 0.02 2.9 8.4 ± 2.6 20 1.7 ± 0.3 4.8 6.8 ± 1.1 20 8.1 ±1.l 0.8 paraformaldehyde + 0.05% glutaraldehyde (final concentrations) and stored deep-frozen until analyzed. Samples were stained with Sytol3 (Molecular Probes lnc) at 2.5 flM tinal concentration. Prokaryotes were detected by their signature in a plot 01' side seatter vs. green fluorescence. The division oftolal prokaryotes into high-NA and low-NA prokaryotes served both as an estimate o[ assel11blage structure and 01' relative activity [(Jaso! and del Giorgio, 2000]. Green fluorescenee was used as al1 estimate 01' prokalyote cell size [Gaso¡ ami del Giorgio, 2000] and eonverted to biomass using the alometric equation of Norland [1993]. 2.5. Prokaryote Physiological State [9] The CTC (5 cyano-2, 3 ditolyl tetrazolium chloride) reduction techniqlle was llsed to estimate the physiological state of the prokaryote assemblages. We added 5 mM eTC of a daily-prepared batch to water samples that were incubated (&om 3 to 12 hours) at in situ temperatures. After incubation, the samples were analyzed on board the ship with the cytometer, to measure the red fluorescence of the formazan granules and lhe relative size 01' each formazan particle. 2.6. Prokaryote Production (P) [lO] P was estimated from the incorporation of tritiated leucille (Leu) and thymidine (TdR), following standard proeedures. We used 4 replicates plus two TCA-killed blanks and incubated with 40 nM TdR and 40 nM Leu for 3 to 8 hours. Precipitation was done with ice-cold TCA. lncorporation rates were transformed to earbon production rates with conversion factors of 1.5 kg e mol Leu -1, which assumes no intracellular isotope dilution, and 20 kg C mol TdR -1, derived trom a TdR conversion factor of 1.6 x 10 18 cells mol1 and 13 fg C celr-I. The average value uf TdR and Leu production for each sample was used as the final P estímate. 3. Results and Discussion [11 J The abundance of prokaryotes in the mesopelagic zone 01' the subtropical N E Atlantic averaged 2 10 8 cells r 1, with an average Len and TdR-based productio/l (P) of 33 nmol C 1 I dI, about 7 and 12 fold lower, respectively, than in the epipelagic zone (Table 1). Highest P values were fmmd in the area affected by the African coastal upwelling, while lower values were observed towards the central Atlantic. However, the average turnover of the mesopelagic prokaryote assemblage (0.13 ± 0.02 dI, Table 1) was only 3 times less than that of the epipelagic prokaIyotes, indicative of an aetively growing mesopelagic assemblage. This was eonsistent with the relatively high contribution of prokaryotes with high-nucleie acid (NA) content to the 2 of 4 L03608 ARÍSTE(,UI ET AL.: MESOPELAGIC PROKARYOTES IN THE NE ATLANTIC L03608 164.5 o 164 E ..:; e ·2 163.5 1! e '" " e 163 O " e '" Ol $( 162.5 O _-e . - ...... 10 20 ". I I I I I I I I I I I I 30 40 Elapsed time (h) 4.5 /-e 4 -U / / O / / ···f '" / / Il> / 3.5 -< / O iD 3 rr O 3 Il> 2.5 "' "' -U 0.8 a A Il> -< o 0.6 m " (3 c. '2 (Q 2 O 0.4 Ú o· :J ~ 0.2 '2 1.5 3 o (i 50 60 70 o ~ Figure 1. Time-series experiments. Time variation of oxygel1 consumptiol1, accumulated prokaryote production ami biomass during a respiration experiment (average values ± SE), with an unconcentrated community sampled al 600 m depth from an oceanic station at 26°N, 26°W. assemblage (58 ± 2%, Table 1). lndeed, 8 ± 1% of the mesopclagic prokaryotes were activcly respiring, as indicated by their reduction of the tetrazolium salt ere (Table 1), a fi'action similar to that in the upper ocean (7 ± 1 %). Furthermore, the respiring cells at mesopelagic depths produced 28'?"o larger formazan granules than the surface prokaryotes (t-test, P < 0.(5), which suggests that the respiration rate per cell was higher in the mesopelagic zone. [12] Prokaryote abundance increased upon confinement of the samples for respiration experiments with rates averaging 0.44 ± 0.04 dI (average doubling time: 2.8 days). The 3-days time-series experiment showed that the oxygen concentration decreased and the cumulative heterotrophic production increased gradually along the experiment (Figure 1). The integrated prokaryote respiration ÜllTIol O2 1-1) throughout all the experiments was dosely related to the integrated prokaryote production (fimo 1 e 11 ), measured both as the change in biomass (r = 0.81, n = 30) and frol11 the average Leu and TdR uptake (r = 0.79, n = 30) (Figure 2). The measured daily rate of prokaryote respiration (R) was also c\osely relatcd to the average abundance ofprokaryotes (PA), during each experiment (Figure 3), in spite that the average eell-specific respiration rate apparently deereased towards the end of the time-series experil11ent. R was therefore corrected using the re\ationship with PA obtained (R = 7.2 10 11 PAIK7lO17; Figure 3) to aceount for the observed inerease in prokaryote abundanee. Using this equation, we derivcd a value of R corresponding to the prokaryote assemblage representative of conditions in situ (i .e., the prokaryote abundan ce in si de the Niskin bottles). The resulting corrected estimates of R averaged 0.22 ± 0.05 f-Lmol O2 I 1 d· -I across the study area, being about 4 fold higher at 600 111 (O J S ± 0.08 fLmol O2 lid 1) than at 1000 m depth (0.08 ± 0.03 fllll01 O2 11 d1 ). These rates are about two thirds of those directly derived from the incubation of uneoneentrated samples (0.33 ± 0.03 fImol O2 1-1 dI ), showing that the re-scaling procedure probably accounted for potential artifacts derived frol11 growth atier conftnement. [13] The derived respiration rates for the rnesopelagie subtropical NE Atlantic are about one order of magnitue!e -N o -o 10 E E, e o ~ '0. (/) CIl a: -. change in biomass --o-- average leucine and Ihymidine • • O O o c9 O O o Prokaryote heterotrophic production (¡¡mole ¡-') Figure 2. Tbe relationship between integrated prokaryote respiration and integrated prokaryote production across all the experiments, with concentrated ane! unconcentrated water samples. The production was estimated both as the prokaryote biol11ass increment in respiration experiments and as the average tritiatee! leueine and thymidine uptake. higher than the average global rate implied in previous assessments (0.02 flmol O2 1-1 dI) [Arístegui et al., 2003]. The reason tar this discrepancy l11ay be two-fold: (l) the subtropical NE Atlantic is an area with particularIy high respiratory activity, as it reeeives substantial lateral organic inputs frOl11 the African upwelling which may fue! mesopelagic R; (2) previous estil11ates of the global average mesopelagic R may be underestimates as these were indirect ones, largely infen'ed trom measured ETS activity and a R/ETS of 0.09, derived from senescent bacterial cultures [Packard el al., 1988; Arístegui ef al., 2003]. However, the results presented here depict the mesopelagic prokaryote 10 '-o -N O -O E E, CIl ~ e o .~ '0. (/) Ql a: 0.1 O 00 O O O o O O O O O O O O O O O o lOS 10 6 Average prokaryote abundance (cells mi") Figure 3. The rclationship between prokaryote respiration rate (R) and the average prokaryote abundance (PA) across aH the experiments, with concentrated and unconcentrated water samplcs. PA was measured as the average value among initial and final concentratiolls of prokaryotes inside the experimental bottlcs. The solíd line represents the fitted regression (Model ll, reduced major axis) equation, R = 7.2 1011 PA I87 ±ÜI7 (r2 == 0.76, P < 0.0(1). :; of 4 L03608 ARÍSTEGUl ET AL.: MESOPELAGIC PROK.A.RYOTES IN THE NE ATLANTlC L03608 assemblages as actively growing (Table 1) rather than senescent. Indeed, the average RJETS ratio in our data set, afier exc1uding two data points from two coastal stations in the Cape Blanc upweIling, was 0.68 (SE ± 0.11; 0,9 without exc1uding Ihe two outlier points), 8 fold greater Ihan that assumed in the past. The rather high RJETS ratio in the mesopelagic subtropical NE Atlantic is similar to that of the mixed layer microbial assemblages [Arísteglli and Montero, 1995]. More interestingly, the ratio is comparable to that (average RIETS = 1.1, range = 0.6-1.7) observed during the exponential growth phase of the same bacterial cultures used lo derive the R/ETS in their senescent state [Christensen et al., 1980]. Our findings suggest that the general use ofa low ratio «0.1) to derive mesopelagic R from ETS may lead to gross underestimation 01' this proccss in the global ocean, i1' prokaryote assemblages are actively growing as observed in our study. Indeed, the prokaryote growth effíciency [PGE = P/(P + R)], estimated with the back-scaled R and the Leu/TdR-based r was, on average, 0.18 ± 0.03 al 600 m, and 0.13 ± 0.02 at 1000 m. These PGE values are comparable to the average PGE of sea surface prokaryote assemblages [del Giorgio and Cole, 2000], and suggest that mesopelagic R mus! be comparable to that of the epipelagic zone if similar integrated abundan ces of prokaryotes are found in both zones, as was the case in our region of study. [14] We further calculated integrated mesopelagie R in the subtropical NE Atlantic by combining 10 vertical profiles (200-1000 m) ofETS activity with the empirically-estimated R to ETS ratio of 0.68, yielding an average value of 68 ± 8 mmol C m2 d1 (assuming a respiratolY quotient, RQ = 0.69). The mesopelagic R is comparable to the epipela~ic R measured during the same cruise (64-97 mmol C m - dI (N. NavatTO, personal communication, 2004» or in previous studies (89-136 mmol C m~2 dI) in the same region [Duarte el al., 2001]. This indicates a very high carbon demand by mesopelagic prokaryotcs that cannot possibly be fulfilled by vertical inpuls of organic carbon alone. EsíÍmates of vertical inputs 01' organic carbon into the mesopelagic layer at this region are well below these values (unpublished results), and tlle activity of the layer must be, therefore, largely fucled by lateral exports from the highly produetive NW African upwelling system. [15] Several sets of evidence support this hypothcsis. Recent studies in ihe NW African coast indicate that upwelling filaments may transport offshore up to 50% of coastal upwelling primary produetion, even during low to modera te upwelling pulses [Gabric et al., 1993; Gal'GÍaMuñoz el al., 2004]. Upwelling filaments are known to be a widespread and recurrent phenomena in the NW AÜ'ican coast [Kostianoy and Zatsepin, 1996]; hence their annual contribution to the coastal-otfshore exchange of organic matter must be impOliant. Additionally, high concentrations of particulate organic carbon (POC) might be aIso transported to the open oeean via intennediate or deep nepheloid layers. Indeed, vertical protiles of suspended roc from the same cruise (unpublished data) show carbon eoncentrations 3-6 ¡.LM higher in the mesopelagic zone of the region of study compared to published results from the North Atlantie subtropical gyre. Moreover, the protiles show peaks ofPOC in the upper 1000 m, more intense towards the coasi, which could only be explaincd by lateral advection of organic malter from the African shelf. [](í] In conclusion, our findings eonfirm that mesope1agic prokatyotc assemblages are active nodes of organic carbon remineralisations and ae! as major sinks for organic carbon in the subtropical N E Atlantic Ocean. A large par! of the mcsope1agic respiration must be supported by coastal adveclÍon of organic matter from the highly productive upwelling system to the open ocean. The possibility that the high mesopelagic R reported here could be a [eature of other eastem boundary regions, bordering high produetive upwelling systems, awaits verification. If confinned, the overall organic carbon sink in these regions would playa key role in lhe global earbon cycle in the oceans. [17] Acknowlcdgmcnts. This research is part 01' the COCA project. fundcd by the Spanish Plan Nacional de I+D (REN2000 1471-C02-01MAR). LA and JMG were also supported by pwjects MICRODIFF (REN2001-2120/MAR) and DEBACOCA (REN2001-4211-E). Hugh Ducklow. Pe ter LeB. Williams and ooe anonymous rcvicwer providcd valuable comments. which helped improving the manuscript. References Arístegui, J.. and M. F. Montero (1995). The relationship between community respiration and ETS activity in [he oeean . .! Plallkroll Res .. 17, 15631571. Arístegui, J., S. AguslÍ, and C. M. Duarte (2003 J. Respiration in !he dark ocean, GCOph\':5. Res. Letr., 30(2), 1041, doi:10.1029!2002GLOI6227. Cho. B. c.. i1nd F. Azall1 (1988). Major role of bacteria in biogcochemical fluxes in the ocean's interior. Nature, 332,441-443. Christensen, 1. P.. T. G. Owens, A. H. Devol. and T. T. Packard (1980). Respiration and physiological state in marine bacteria. Mw: Sio/., 55. 267·,,·276. del Giorgio. P. A .• and J . .1. Cole (2000), Bacterial cncrgclics and growth efficieney, in Microbial Ec%gy o(the Occans. edited by D. L. Kirchman. pp. 289326. John Wiley. Hoboken. N. 1. del Giorgio, P. A., and C. M. Duarte (2002), Respiration in the open orean, Nat",.", 420. 379-384. Duarte, C. M., S. Agustí, J. Arístegui. N. González, and R. Anadón (2001 1, Evidence for a heterotrophie subtropical northeast Atlantic. Lillll1ol. Oceanogr., 46, 425-428. Gabríe, A. J .. L. Garda. l. van Campo L. Nykjaer. W. Eiller. and W. Schrimpf (1993). Offshore export of shelf production in the Cape Blanc (Maurítania) giant tílamen! as derivcd fmm coastal zone color scanner imagery,.J. Geophys. Res .. 98.46974712. García-Muñoz, M .• J. Aristegui, M. F. Montero. anel E. D. Barton (2004). Distribution anel transport of organic matter along a tílament-eddy system in the CanariesNW Ají'ica coas!al transition zone region, Prog. Oceanogr., 62, 115-129. Gasol, 1. M., and P. A. del Giorgío (2000), Using flow cytometry fOf counting natural planktonic bacteria and understaneling the stnlcturc 01' planktonic bacterial commllnities, Sd Mm:, 64. 197224. Jenkins, W. J. (1977). Tritium-helillm dating in the Sargasso Sea: A 111casurcll1enl of oxygen utilization rates. Science. /96.291-292. Jcnkins. W. J. (19821, Oxygen utilization rutes in lhe North Atlantic Subtropical Gyrc and primary produetion in oligotrophic systems. Na/l/re. 300. 246-·248. Kostianoy. A. G., and A. G. Zatsepin (1996). The \Vest Ateican coastal upwelling tilall1ents and cross-thmtal water exchange conditioned by them, 1. Ma~ Svsl., 7,349-359. Norland. S. (1993). The rclationship betwccn biomass and volumc oí" bacteria, ín Halldbook oj"Methods in Aqllatic Microbia! Ecology, cdited by P. F. Kemp et al., pp. 303 307, Lewis, Boca Raton, Fla. Packard, T. T., M. Denis, M. Rodier. and P. Garfield (19881. Deep-ocean mctabolie CO 2 production: Calculations ¡¡'om ETS activity, De,'l' Sea Res .. Part 1.35.371-382. J. Arístegui, Departamento de Biología. Universidad de las Palmas de Gran Canaria, Las Palmas dc Gran Canaria, E-350 I 7 Las Palmas de Gran Canaria. Islas Canarias, Spain. ([email protected]) L. Alonso-Sáez and J. M. Gasol. Departament de Biologia Marina i Oceanografia, Institut de Ciéncies del Mar-CMIMA, ('SIC. Pg Marítim de la Barceloncta 37-49, E-OR003 Barcelona. Catallwya. Spain. C. M. Duartc, IMEDEA, CSIC-UiB, CiMiqucl Marques 21. E-07190 Esporles, Islas Baleares, Spain. 4 af 4